Predictive and Prescriptive Analysis.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## This Notebook covers predictive data analysis part\n",
    "\n",
    "Predictive: What is likely to happen?\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/dulanj/Environments/py376/lib/python3.7/site-packages/sklearn/externals/six.py:31: FutureWarning: The module is deprecated in version 0.21 and will be removed in version 0.23 since we've dropped support for Python 2.7. Please rely on the official version of six (https://pypi.org/project/six/).\n",
      "  \"(https://pypi.org/project/six/).\", FutureWarning)\n",
      "Using TensorFlow backend.\n",
      "/home/dulanj/Environments/py376/lib/python3.7/site-packages/tensorflow/python/framework/dtypes.py:526: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_qint8 = np.dtype([(\"qint8\", np.int8, 1)])\n",
      "/home/dulanj/Environments/py376/lib/python3.7/site-packages/tensorflow/python/framework/dtypes.py:527: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_quint8 = np.dtype([(\"quint8\", np.uint8, 1)])\n",
      "/home/dulanj/Environments/py376/lib/python3.7/site-packages/tensorflow/python/framework/dtypes.py:528: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_qint16 = np.dtype([(\"qint16\", np.int16, 1)])\n",
      "/home/dulanj/Environments/py376/lib/python3.7/site-packages/tensorflow/python/framework/dtypes.py:529: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_quint16 = np.dtype([(\"quint16\", np.uint16, 1)])\n",
      "/home/dulanj/Environments/py376/lib/python3.7/site-packages/tensorflow/python/framework/dtypes.py:530: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_qint32 = np.dtype([(\"qint32\", np.int32, 1)])\n",
      "/home/dulanj/Environments/py376/lib/python3.7/site-packages/tensorflow/python/framework/dtypes.py:535: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  np_resource = np.dtype([(\"resource\", np.ubyte, 1)])\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import seaborn as sns\n",
    "import numpy as np\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.tree import DecisionTreeRegressor\n",
    "from sklearn.metrics import r2_score\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "import pydotplus\n",
    "import matplotlib.image as mpimg\n",
    "import io\n",
    "\n",
    "from sklearn.linear_model import LinearRegression\n",
    "from sklearn.model_selection import RandomizedSearchCV\n",
    "from pprint import pprint\n",
    "\n",
    "from sklearn.externals.six import StringIO\n",
    "from sklearn.tree import export_graphviz\n",
    "\n",
    "from sklearn.svm import SVR\n",
    "from sklearn.ensemble import RandomForestRegressor\n",
    "\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Load data from excel file"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Index(['Date', 'Product', 'Planned start Time', 'Actual Start Time',\n",
      "       'Stop Time', 'Total time', 'Startup Delay', 'BreakDowns',\n",
      "       'Eff Prod time', 'Operating Speed', 'Produced Qty', 'Rejects Qty',\n",
      "       'Mix Usage', 'Mix Wastage', 'Mix yield', 'AR', 'PR', 'QR', 'OEE',\n",
      "       'Mix Per Stick', 'Efficiency', 'Planned Qty'],\n",
      "      dtype='object')\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Date</th>\n",
       "      <th>Product</th>\n",
       "      <th>Planned start Time</th>\n",
       "      <th>Actual Start Time</th>\n",
       "      <th>Stop Time</th>\n",
       "      <th>Total time</th>\n",
       "      <th>Startup Delay</th>\n",
       "      <th>BreakDowns</th>\n",
       "      <th>Eff Prod time</th>\n",
       "      <th>Operating Speed</th>\n",
       "      <th>...</th>\n",
       "      <th>Mix Usage</th>\n",
       "      <th>Mix Wastage</th>\n",
       "      <th>Mix yield</th>\n",
       "      <th>AR</th>\n",
       "      <th>PR</th>\n",
       "      <th>QR</th>\n",
       "      <th>OEE</th>\n",
       "      <th>Mix Per Stick</th>\n",
       "      <th>Efficiency</th>\n",
       "      <th>Planned Qty</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2018-07-02</td>\n",
       "      <td>WB Chocolate</td>\n",
       "      <td>06:30:00</td>\n",
       "      <td>10:30:00</td>\n",
       "      <td>20:30:00</td>\n",
       "      <td>14.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>5.15</td>\n",
       "      <td>4.85</td>\n",
       "      <td>28880</td>\n",
       "      <td>...</td>\n",
       "      <td>7840.0</td>\n",
       "      <td>991.000</td>\n",
       "      <td>0.873597</td>\n",
       "      <td>0.346429</td>\n",
       "      <td>1.496754</td>\n",
       "      <td>0.873597</td>\n",
       "      <td>0.452976</td>\n",
       "      <td>0.0375</td>\n",
       "      <td>0.499675</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2018-07-04</td>\n",
       "      <td>WB Vanilla</td>\n",
       "      <td>06:00:00</td>\n",
       "      <td>18:30:00</td>\n",
       "      <td>20:00:00</td>\n",
       "      <td>14.0</td>\n",
       "      <td>12.5</td>\n",
       "      <td>0.00</td>\n",
       "      <td>1.50</td>\n",
       "      <td>28880</td>\n",
       "      <td>...</td>\n",
       "      <td>1160.0</td>\n",
       "      <td>40.625</td>\n",
       "      <td>0.964978</td>\n",
       "      <td>0.107143</td>\n",
       "      <td>0.716049</td>\n",
       "      <td>0.964978</td>\n",
       "      <td>0.074033</td>\n",
       "      <td>0.0375</td>\n",
       "      <td>0.098710</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2018-07-05</td>\n",
       "      <td>WB Vanilla</td>\n",
       "      <td>06:00:00</td>\n",
       "      <td>09:00:00</td>\n",
       "      <td>18:00:00</td>\n",
       "      <td>12.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>3.90</td>\n",
       "      <td>5.10</td>\n",
       "      <td>28880</td>\n",
       "      <td>...</td>\n",
       "      <td>7840.0</td>\n",
       "      <td>812.125</td>\n",
       "      <td>0.896413</td>\n",
       "      <td>0.425000</td>\n",
       "      <td>1.423384</td>\n",
       "      <td>0.896413</td>\n",
       "      <td>0.542274</td>\n",
       "      <td>0.0375</td>\n",
       "      <td>0.595438</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2018-07-06</td>\n",
       "      <td>JJ Necto</td>\n",
       "      <td>06:00:00</td>\n",
       "      <td>09:00:00</td>\n",
       "      <td>21:00:00</td>\n",
       "      <td>15.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>5.18</td>\n",
       "      <td>6.82</td>\n",
       "      <td>28880</td>\n",
       "      <td>...</td>\n",
       "      <td>16000.0</td>\n",
       "      <td>2413.600</td>\n",
       "      <td>0.849150</td>\n",
       "      <td>0.454667</td>\n",
       "      <td>1.357663</td>\n",
       "      <td>0.849150</td>\n",
       "      <td>0.524167</td>\n",
       "      <td>0.0600</td>\n",
       "      <td>0.562595</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2018-07-07</td>\n",
       "      <td>WB Vanilla</td>\n",
       "      <td>06:00:00</td>\n",
       "      <td>10:30:00</td>\n",
       "      <td>18:00:00</td>\n",
       "      <td>12.0</td>\n",
       "      <td>4.5</td>\n",
       "      <td>5.10</td>\n",
       "      <td>2.40</td>\n",
       "      <td>28880</td>\n",
       "      <td>...</td>\n",
       "      <td>7000.0</td>\n",
       "      <td>508.750</td>\n",
       "      <td>0.927321</td>\n",
       "      <td>0.200000</td>\n",
       "      <td>2.700617</td>\n",
       "      <td>0.927321</td>\n",
       "      <td>0.500868</td>\n",
       "      <td>0.0375</td>\n",
       "      <td>0.605216</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 22 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "        Date       Product Planned start Time Actual Start Time Stop Time  \\\n",
       "0 2018-07-02  WB Chocolate           06:30:00          10:30:00  20:30:00   \n",
       "1 2018-07-04    WB Vanilla           06:00:00          18:30:00  20:00:00   \n",
       "2 2018-07-05    WB Vanilla           06:00:00          09:00:00  18:00:00   \n",
       "3 2018-07-06      JJ Necto           06:00:00          09:00:00  21:00:00   \n",
       "4 2018-07-07    WB Vanilla           06:00:00          10:30:00  18:00:00   \n",
       "\n",
       "   Total time  Startup Delay  BreakDowns  Eff Prod time  Operating Speed  ...  \\\n",
       "0        14.0            4.0        5.15           4.85            28880  ...   \n",
       "1        14.0           12.5        0.00           1.50            28880  ...   \n",
       "2        12.0            3.0        3.90           5.10            28880  ...   \n",
       "3        15.0            3.0        5.18           6.82            28880  ...   \n",
       "4        12.0            4.5        5.10           2.40            28880  ...   \n",
       "\n",
       "   Mix Usage  Mix Wastage  Mix yield        AR        PR        QR       OEE  \\\n",
       "0     7840.0      991.000   0.873597  0.346429  1.496754  0.873597  0.452976   \n",
       "1     1160.0       40.625   0.964978  0.107143  0.716049  0.964978  0.074033   \n",
       "2     7840.0      812.125   0.896413  0.425000  1.423384  0.896413  0.542274   \n",
       "3    16000.0     2413.600   0.849150  0.454667  1.357663  0.849150  0.524167   \n",
       "4     7000.0      508.750   0.927321  0.200000  2.700617  0.927321  0.500868   \n",
       "\n",
       "   Mix Per Stick  Efficiency  Planned Qty  \n",
       "0         0.0375    0.499675          NaN  \n",
       "1         0.0375    0.098710          NaN  \n",
       "2         0.0375    0.595438          NaN  \n",
       "3         0.0600    0.562595          NaN  \n",
       "4         0.0375    0.605216          NaN  \n",
       "\n",
       "[5 rows x 22 columns]"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = pd.read_excel('../data/Production Data Versaline.xlsx')\n",
    "df = df.rename(columns={\"Breakdowns and other delays\": \"BreakDowns\", \"Effective production time\": \"Eff Prod time\"})\n",
    "print(df.columns)\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Calculate Rejects percentage and add as a colomn\n",
    "\n",
    "Rejects percentage = Rejects Qty / (Total Qty + Rejects Qty)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Add rejects percentage\n",
    "def get_rejects_percentage(row):\n",
    "    return row[\"Rejects Qty\"]/(row[\"Produced Qty\"]+row[\"Rejects Qty\"])\n",
    "\n",
    "df['Rejects Pct'] = df.apply (lambda row: get_rejects_percentage(row), axis=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Fill the NAN values with the mean of the colomn"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 441 entries, 0 to 440\n",
      "Data columns (total 23 columns):\n",
      " #   Column              Non-Null Count  Dtype         \n",
      "---  ------              --------------  -----         \n",
      " 0   Date                441 non-null    datetime64[ns]\n",
      " 1   Product             441 non-null    object        \n",
      " 2   Planned start Time  441 non-null    object        \n",
      " 3   Actual Start Time   441 non-null    object        \n",
      " 4   Stop Time           441 non-null    object        \n",
      " 5   Total time          441 non-null    float64       \n",
      " 6   Startup Delay       441 non-null    float64       \n",
      " 7   BreakDowns          441 non-null    float64       \n",
      " 8   Eff Prod time       441 non-null    float64       \n",
      " 9   Operating Speed     441 non-null    int64         \n",
      " 10  Produced Qty        441 non-null    int64         \n",
      " 11  Rejects Qty         440 non-null    float64       \n",
      " 12  Mix Usage           441 non-null    float64       \n",
      " 13  Mix Wastage         441 non-null    float64       \n",
      " 14  Mix yield           439 non-null    float64       \n",
      " 15  AR                  441 non-null    float64       \n",
      " 16  PR                  441 non-null    float64       \n",
      " 17  QR                  441 non-null    float64       \n",
      " 18  OEE                 441 non-null    float64       \n",
      " 19  Mix Per Stick       441 non-null    float64       \n",
      " 20  Efficiency          441 non-null    float64       \n",
      " 21  Planned Qty         73 non-null     float64       \n",
      " 22  Rejects Pct         441 non-null    float64       \n",
      "dtypes: datetime64[ns](1), float64(16), int64(2), object(4)\n",
      "memory usage: 79.4+ KB\n"
     ]
    }
   ],
   "source": [
    "for val in [\"Rejects Pct\"]:#[\"Rejects Pct\", \"Mix yield\"]\n",
    "    df[val] = df[val].fillna(df[val].mean())\n",
    "    \n",
    "df.info()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Select only the useful features\n",
    "remove duplicate features such as start time and end time"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Product</th>\n",
       "      <th>Produced Qty</th>\n",
       "      <th>BreakDowns</th>\n",
       "      <th>Startup Delay</th>\n",
       "      <th>Total time</th>\n",
       "      <th>Rejects Pct</th>\n",
       "      <th>OEE</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>WB Chocolate</td>\n",
       "      <td>182640</td>\n",
       "      <td>5.15</td>\n",
       "      <td>4.0</td>\n",
       "      <td>14.0</td>\n",
       "      <td>0.126403</td>\n",
       "      <td>0.452976</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>WB Vanilla</td>\n",
       "      <td>29850</td>\n",
       "      <td>0.00</td>\n",
       "      <td>12.5</td>\n",
       "      <td>14.0</td>\n",
       "      <td>0.035022</td>\n",
       "      <td>0.074033</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        Product  Produced Qty  BreakDowns  Startup Delay  Total time  \\\n",
       "0  WB Chocolate        182640        5.15            4.0        14.0   \n",
       "1    WB Vanilla         29850        0.00           12.5        14.0   \n",
       "\n",
       "   Rejects Pct       OEE  \n",
       "0     0.126403  0.452976  \n",
       "1     0.035022  0.074033  "
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "useful_colomns = [\"Product\", \"Produced Qty\", \"BreakDowns\", \"Startup Delay\", \"Total time\", \"Rejects Pct\", \"OEE\"]\n",
    "df = df[useful_colomns]\n",
    "df.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['WB Chocolate', 'WB Vanilla', 'JJ Necto', 'MM Watermelon',\n",
       "       'MM Orange', 'WB Chock Shock', 'WB Berry', '2-Bar Vanilla Faluda',\n",
       "       'JJ Cream Soda', 'JJ Orange Crush ', 'Ice Cake', 'Bubble Gumy',\n",
       "       'JJ Creamsoda', 'WB Vnilla', 'Bubble Gummy', 'Ice cake',\n",
       "       'WB Chocok Shock', 'JJ Orange Crush'], dtype=object)"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.Product.unique()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Rename Duplicates products with the correct name"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WB Chocolate    = 62\n",
      "WB Vanilla      = 74\n",
      "JJ Necto        = 45\n",
      "MM Watermelon   = 51\n",
      "MM Orange       = 65\n",
      "WB Chock Shock  = 30\n",
      "WB Berry        = 16\n",
      "2-Bar Vanilla Faluda = 65\n",
      "JJ Cream Soda   = 15\n",
      "JJ Orange Crush = 8\n",
      "Ice Cake        = 5\n",
      "Bubble Gummy    = 5\n"
     ]
    }
   ],
   "source": [
    "df = df.replace({'Product': {\"JJ Orange Crush \": \"JJ Orange Crush\", \n",
    "                             \"WB Chocok Shock\": \"WB Chock Shock\",\n",
    "                            \"Ice cake\": \"Ice Cake\",\n",
    "                            \"Bubble Gumy\":\"Bubble Gummy\", \n",
    "                            \"WB Vnilla\": \"WB Vanilla\",\n",
    "                            \"JJ Creamsoda\": \"JJ Cream Soda\"}})\n",
    "for product in df.Product.unique():\n",
    "    print(\"{:15} = {}\".format(product, df[df[\"Product\"] == product][\"Product\"].count()))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of rows: 441\n",
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 441 entries, 0 to 440\n",
      "Data columns (total 7 columns):\n",
      " #   Column         Non-Null Count  Dtype  \n",
      "---  ------         --------------  -----  \n",
      " 0   Product        441 non-null    object \n",
      " 1   Produced Qty   441 non-null    int64  \n",
      " 2   BreakDowns     441 non-null    float64\n",
      " 3   Startup Delay  441 non-null    float64\n",
      " 4   Total time     441 non-null    float64\n",
      " 5   Rejects Pct    441 non-null    float64\n",
      " 6   OEE            441 non-null    float64\n",
      "dtypes: float64(5), int64(1), object(1)\n",
      "memory usage: 24.2+ KB\n"
     ]
    }
   ],
   "source": [
    "print(\"Number of rows: {}\".format(df.Product.count()))\n",
    "df.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "# group_by_products = df.groupby(by=['Product'])\n",
    "# product_data_avg = group_by_products.mean()\n",
    "# product_data_count = group_by_products.count()\n",
    "# print(product_data_avg)\n",
    "# print(product_data_count)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Product              = 12\n",
      "Produced Qty         = 433\n",
      "BreakDowns           = 261\n",
      "Startup Delay        = 130\n",
      "Total time           = 208\n",
      "Rejects Pct          = 301\n",
      "OEE                  = 439\n"
     ]
    }
   ],
   "source": [
    "for col in df.columns.values:\n",
    "    print(\"{:20} = {}\".format(col, df[col].nunique()))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Here we test different regression methods"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "def test_results(model, features, actual):\n",
    "    y_pred = model.predict([features])\n",
    "    print(\"Result: {}\\nActual: {}\".format(y_pred[0], actual))\n",
    "    print(\"Error:{}\\nSqured Error:{}\".format(y_pred[0]-actual, (y_pred[0]-actual)**2))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Data Preperation"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Onehot encoding"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Products as a input\n",
    "df['Product'] = pd.Categorical(df['Product'])\n",
    "dfDummies = pd.get_dummies(df['Product'], prefix = 'prod')\n",
    "# dfDummies.info()\n",
    "df_p = pd.concat([df, dfDummies], axis=1)\n",
    "# df_p.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Select only one product\n",
    "# df_p = df[df['Product']==\"WB Chocolate\"]\n",
    "# df_p = df[df['Product']==\"MM Orange\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "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>Produced Qty</th>\n",
       "      <th>BreakDowns</th>\n",
       "      <th>Startup Delay</th>\n",
       "      <th>Total time</th>\n",
       "      <th>Rejects Pct</th>\n",
       "      <th>prod_2-Bar Vanilla Faluda</th>\n",
       "      <th>prod_Bubble Gummy</th>\n",
       "      <th>prod_Ice Cake</th>\n",
       "      <th>prod_JJ Cream Soda</th>\n",
       "      <th>prod_JJ Necto</th>\n",
       "      <th>prod_JJ Orange Crush</th>\n",
       "      <th>prod_MM Orange</th>\n",
       "      <th>prod_MM Watermelon</th>\n",
       "      <th>prod_WB Berry</th>\n",
       "      <th>prod_WB Chock Shock</th>\n",
       "      <th>prod_WB Chocolate</th>\n",
       "      <th>prod_WB Vanilla</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>361</th>\n",
       "      <td>1.574481</td>\n",
       "      <td>1.237761</td>\n",
       "      <td>1.205744</td>\n",
       "      <td>2.137366</td>\n",
       "      <td>-0.641294</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>249</th>\n",
       "      <td>-1.034454</td>\n",
       "      <td>-0.818599</td>\n",
       "      <td>0.391848</td>\n",
       "      <td>-1.029216</td>\n",
       "      <td>-0.641294</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>271</th>\n",
       "      <td>-0.886922</td>\n",
       "      <td>-0.727756</td>\n",
       "      <td>2.368452</td>\n",
       "      <td>-0.181025</td>\n",
       "      <td>2.778066</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>434</th>\n",
       "      <td>0.659353</td>\n",
       "      <td>2.385689</td>\n",
       "      <td>0.895688</td>\n",
       "      <td>1.359857</td>\n",
       "      <td>0.071918</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>397</th>\n",
       "      <td>-1.062622</td>\n",
       "      <td>-0.958993</td>\n",
       "      <td>-0.809618</td>\n",
       "      <td>-1.758661</td>\n",
       "      <td>-0.641294</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     Produced Qty  BreakDowns  Startup Delay  Total time  Rejects Pct  \\\n",
       "361      1.574481    1.237761       1.205744    2.137366    -0.641294   \n",
       "249     -1.034454   -0.818599       0.391848   -1.029216    -0.641294   \n",
       "271     -0.886922   -0.727756       2.368452   -0.181025     2.778066   \n",
       "434      0.659353    2.385689       0.895688    1.359857     0.071918   \n",
       "397     -1.062622   -0.958993      -0.809618   -1.758661    -0.641294   \n",
       "\n",
       "     prod_2-Bar Vanilla Faluda  prod_Bubble Gummy  prod_Ice Cake  \\\n",
       "361                          1                  0              0   \n",
       "249                          1                  0              0   \n",
       "271                          0                  0              0   \n",
       "434                          0                  0              0   \n",
       "397                          0                  0              0   \n",
       "\n",
       "     prod_JJ Cream Soda  prod_JJ Necto  prod_JJ Orange Crush  prod_MM Orange  \\\n",
       "361                   0              0                     0               0   \n",
       "249                   0              0                     0               0   \n",
       "271                   0              0                     0               0   \n",
       "434                   0              0                     0               0   \n",
       "397                   0              0                     0               0   \n",
       "\n",
       "     prod_MM Watermelon  prod_WB Berry  prod_WB Chock Shock  \\\n",
       "361                   0              0                    0   \n",
       "249                   0              0                    0   \n",
       "271                   0              0                    1   \n",
       "434                   0              0                    1   \n",
       "397                   0              0                    0   \n",
       "\n",
       "     prod_WB Chocolate  prod_WB Vanilla  \n",
       "361                  0                0  \n",
       "249                  0                0  \n",
       "271                  0                0  \n",
       "434                  0                0  \n",
       "397                  0                1  "
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Scale the data\n",
    "y = df_p[\"OEE\"]\n",
    "X = df_p.drop(['Product', 'OEE'], axis = 1) \n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3, random_state=0)\n",
    "\n",
    "def scale(dataframe, scaler=None):\n",
    "    colomns_to_scale = [\"Produced Qty\", \"BreakDowns\", \"Startup Delay\", \"Total time\", \"Rejects Pct\"]\n",
    "    # Only scale relevenat features\n",
    "    scaled_features = dataframe.copy()\n",
    "    features = scaled_features[colomns_to_scale]\n",
    "    if scaler is None:\n",
    "        scaler = StandardScaler().fit(features.values)\n",
    "    features = scaler.transform(features.values)\n",
    "    scaled_features[colomns_to_scale] = features\n",
    "    return scaled_features, scaler\n",
    "\n",
    "X_train_s, scaler = scale(X_train)\n",
    "X_test_s, _ = scale(X_test, scaler)\n",
    "# X_train_s.head()\n",
    "X_test_s.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Linear Regression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.8448161466480698\n"
     ]
    }
   ],
   "source": [
    "regr = LinearRegression() \n",
    "regr.fit(X_train, y_train) \n",
    "y_pred = regr.predict(X_test)\n",
    "print(\"Accuracy:\", r2_score(y_test, y_pred))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Decision Tree "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "# y = df_p[\"OEE\"]\n",
    "# X = df_p.drop(['Product', 'OEE'], axis = 1) \n",
    "\n",
    "# X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3, random_state=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'criterion': ['mse', 'friedman_mse', 'mae'],\n",
      " 'max_depth': [10, 20, 30, 40, 50, 60, 70, 80, 90, 100, 110, None],\n",
      " 'max_features': ['auto', 'log2', 'sqrt', None],\n",
      " 'min_samples_leaf': [1, 2, 4],\n",
      " 'min_samples_split': [2, 5, 10],\n",
      " 'splitter': ['best', 'random']}\n",
      "dict_keys(['ccp_alpha', 'criterion', 'max_depth', 'max_features', 'max_leaf_nodes', 'min_impurity_decrease', 'min_impurity_split', 'min_samples_leaf', 'min_samples_split', 'min_weight_fraction_leaf', 'presort', 'random_state', 'splitter'])\n",
      "Fitting 3 folds for each of 100 candidates, totalling 300 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 8 concurrent workers.\n",
      "[Parallel(n_jobs=-1)]: Done  25 tasks      | elapsed:    0.9s\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'splitter': 'best', 'min_samples_split': 5, 'min_samples_leaf': 2, 'max_features': 'auto', 'max_depth': 10, 'criterion': 'mae'}\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Done 300 out of 300 | elapsed:    1.2s finished\n"
     ]
    }
   ],
   "source": [
    "# Hyper parameter Tuning for Decision Trees\n",
    "# Number of features to consider at every split\n",
    "criterion = [\"mse\", \"friedman_mse\", \"mae\"]\n",
    "splitter = [\"best\", \"random\"]\n",
    "max_features = ['auto', 'log2', 'sqrt']\n",
    "max_features.append(None)\n",
    "# Maximum number of levels in tree\n",
    "max_depth = [int(x) for x in np.linspace(10, 110, num=11)]\n",
    "max_depth.append(None)\n",
    "# Minimum number of samples required to split a node\n",
    "min_samples_split = [2, 5, 10]\n",
    "# Minimum number of samples required at each leaf node\n",
    "min_samples_leaf = [1, 2, 4]\n",
    "# Method of selecting samples for training each tree\n",
    "bootstrap = [True, False]  # Create the random grid\n",
    "random_grid = {'criterion': criterion,\n",
    "               'splitter': splitter,\n",
    "                'max_features': max_features,\n",
    "               'max_depth': max_depth,\n",
    "               'min_samples_split': min_samples_split,\n",
    "               'min_samples_leaf': min_samples_leaf}\n",
    "pprint(random_grid)\n",
    "\n",
    "# Use the random grid to search for best hyperparameters\n",
    "# First create the base model to tune\n",
    "rf = DecisionTreeRegressor()\n",
    "print(rf.get_params().keys())\n",
    "# Random search of parameters, using 3 fold cross validation,\n",
    "# search across 100 different combinations, and use all available cores\n",
    "rf_random = RandomizedSearchCV(estimator=rf, param_distributions=random_grid, n_iter=100, cv=3, verbose=2,\n",
    "                               random_state=42, n_jobs=-1)  # Fit the random search model\n",
    "rf_random.fit(X_train, y_train)\n",
    "print(rf_random.best_params_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "DecisionTreeRegressor(ccp_alpha=0.0, criterion='mae', max_depth=50,\n",
      "                      max_features=None, max_leaf_nodes=None,\n",
      "                      min_impurity_decrease=0.0, min_impurity_split=None,\n",
      "                      min_samples_leaf=2, min_samples_split=5,\n",
      "                      min_weight_fraction_leaf=0.0, presort='deprecated',\n",
      "                      random_state=0, splitter='best')\n",
      "Accuracy: 0.8343317707950609\n"
     ]
    }
   ],
   "source": [
    "regressor = DecisionTreeRegressor(splitter=\"best\", criterion=\"mae\", random_state=0, min_samples_split=5, \n",
    "                                  min_samples_leaf=2, max_features=None, max_depth=50)\n",
    "# regressor = DecisionTreeRegressor(random_state=0)\n",
    "print(regressor.fit(X_train,y_train))\n",
    "\n",
    "y_pred = regressor.predict(X_test)\n",
    "print(\"Accuracy:\", r2_score(y_test, y_pred))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Result: 0.325280898876404\n",
      "Actual: 0.355357142857143\n",
      "Error:-0.030076243980738993\n",
      "Squred Error:0.0009045804519889385\n"
     ]
    }
   ],
   "source": [
    "number = 2\n",
    "test_results(regressor, X_test.iloc[number], y_test.iloc[number])\n",
    "# y_pred = regressor.predict([X.loc[number]])\n",
    "# print(\"Features: {}\\nResult: {}\".format(X.loc[number], y_pred))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Plotting the tree"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn import tree\n",
    "t = tree.plot_tree(regressor)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['Produced Qty', 'BreakDowns', 'Startup Delay', 'Total time', 'Rejects Pct', 'prod_2-Bar Vanilla Faluda', 'prod_Bubble Gummy', 'prod_Ice Cake', 'prod_JJ Cream Soda', 'prod_JJ Necto', 'prod_JJ Orange Crush', 'prod_MM Orange', 'prod_MM Watermelon', 'prod_WB Berry', 'prod_WB Chock Shock', 'prod_WB Chocolate', 'prod_WB Vanilla']\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAswAAAClCAYAAABItauhAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjMsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+AADFEAAAgAElEQVR4nOyddVgUXRvG7wEFFKUFxQ5Exdfu+AxAXxVsUVSwsbuw+1WxUAwwERsxwQRMbBELFQMb6W7B+f4YdnaX7WIBz++69tqZM+fMPBszc885z3keiqZpEAgEAoFAIBAIBOFoqNsAAoFAIBAIBAKhOEMEM4FAIBAIBAKBIAYimAkEAoFAIBAIBDEQwUwgEAgEAoFAIIiBCGYCgUAgEAgEAkEMRDATCAQCgUAgEAhiKHLBTFHUvxRFRVAU9ZGiKNeiPj6BQCAQCAQCgSALVFHGYaYoShPAewC2AH4AeALAkabpN0VmBIFAIBAIBAKBIANF3cPcBsBHmqYjaZrOBXASQL8itoFAIBAIBAKBQJCaohbMVQF851n/UVBGIBAIBAKBQCAUS8qo24DCUBTlAsAFAHR1dVs2aNBAzRYRCAQCQZ1ERrxBTTMDhfeTmpEDw6q1lWARgUAojYSGhsbTNF1J2LaiFsw/AVTnWa9WUMZC0/ReAHsBoFWrVvTTp0+LzjoCgUAgFDuSzq7AnoBHiE5Mg8eF+0g8sxxGg1Yj8cxygbpGg1ajUU1ThGydJLSO4cBVRWU2gUAoYVAU9VXUtqIWzE8AWFAUVRuMUB4GYHgR20AgEAiEEsZku7YAgFXONgCAdWN6AgCO3XiOEd2bsfU4AtnUYa1QQU0gEAjyUKSCmabpPIqipgG4BkATwEGapsOL0gYCgUAglHw4AppXLPMS67u0KM0hEAilnCKPw0zT9GWapuvTNF2Xpul1RX18AoFAIBRf7O3t4ezsLLGe393XABgXDABoP3MPu+1sCOmHIRAIyoVk+iMQCASC0unUqRNmzpwpczt/f3/4+PhIrOfifhYAcGShAwDgwfbJrHge2MkKiw9ek/qYEydORH5+vsy2EgiEv4ciTVwiK2TSH4FAIKgPc3NzLFy4UC7hq0w0NDSQ4LdM4f3UH7cNcUmpMrf7/fs3pk2bBi8vL4VtIBAIxReKokJpmm4ldBsRzAQCgVC60dTUxKFDh6RydSDIzt69e9GoUSN06tRJ3aYQCAQFECeYi10cZgKBQCDwExUVhapVq+LixYuwt7eXuT1xN1AtLi4uEuvUrVsXnz59KgJrCASCKiA+zAQCgVBEhIeHg6IohIfLNinN3NwcNE3LJZYJxQNxYjkyMhIbN24sQmsIBIKsEJcMAoFQLMh4coZdrtBmMNIf+wmtF/ggDMPmu+HLtYOo0s0J6Y/9oNt6EPLz86GpqalyO0NCQtC5c2f8/PkT5ubmKj8egQAAFEVBlvv1oiEdQIHC3P5tBLaZjHRH/NFZfGU3Xn5F9yY1cT3sM3o0rw3zMbuQk5ursN0EQkmC+DATCIRiD69glhXd1oNkbuPj44NRo0bhz58/oChK7mMTCMUBBwcH+Pr6susJx2bzbS8skt3OPsSCge3E7tN4xDblGkkgFHPECWbikkEgEIodgQ/C0HrYbBwNuMmWJSSnAQAWbD2EZ2/5h7cHDZJdMDs7O4OmaSKWCaUCXrEsjCm9WyI1MweBzz8DgESxTCAQ+CGCmUAgFDts2zdHo7o1MNKuGyq0GYylHkdgbFARALD75CV0HePKV//MGfl7pwmE0o7JSHfsvhyKbRefwHHzBZiMdMf+wBcwGekOAMj/Q+Pk3TdqtpJAKN6QKBkEAkEp2NvbIykpCSEhIUrZ3+F1zJByYV/m9Md+2HUiQCnHIBD+BnhdMVYM44a+u/DoPY7dDseK43fwdvdEsfswMjJCYmKiymwkEIo7xIeZQCAAUL7glZWi9mEmEEozhX2YZWXnpVCsOH5HpjayTkwkEIobZNIfgVDKsbKyQqdOnUgmMgKBUKzZuHEj6tSpgyFDhqjbFAJBACKYCYRiDhG8BAKBwGBvbw8vLy8StpFQ5JAoGQSCioiKioKmpiZWrlyp0H7Cw8OJWBZD165d1W0CgVCqmDx5srpNEIm/v79UYpmkIicUJUQwE0RCUZRCL1sba3V/BIlERkaiTJky2L59u1ztzc3NkZ+fr7Bg/tuR9F+6ffu2xDoTJ4qftEQg/G0MHjRI5Pni6ekpcltJQdJ8i6CgILi6uoqtIwuK3hNL2vdL4Ie4ZBCEkhq4E/o9pgMAdsx2xKheHaDfYzpSrnuIbKPfYzqee69AbXMTAIDd/B2wam+NPXv2qMRGTsa1w4cPw9nZWSXHIKie7A/3+dart+uDhKQU0DSNrPf3RLYrV78jbp3yQtvmjQEA/zpPx60HoSq1lUAoKQQFBcHW1hYaFAWKAn55jsK8o/fxOTYNZ+b05Kv75FMsWtc1hZ3bZYzu0gCD29ZBpQmH1GR50SPNZMWks8v51i89joDTRl8knlkGo0FrAACJZ5aBpoHCmrjn4kN4EvEDnRvXwoVVTgAAw4GrlfcBCEqD+DATZCY1cKdS9jN+3x2RAfWJ4CUAgoJZFAlJKegwcCyWzRgP646tUcXURKCOjkUHZZtHIJRIfH190S3lisR6DeacxLutwwTK/ybBLAmKopB4ZplUdXPz8qFVRhMAYOt6EIEbxgqtRwRz8UScYCZxmAlqo1OnTiQEEUFqjA31EXGTJCghEJSJMLFM4IemaYEeZlFwxDIAkWKZUDIhPswEqRm8hHGtCHryBuuPXMbBAH7/sYysHHWYRSiFZGXnoFz9jgCANvajkPv7N9/2+l0HqsMsAqFEcvh2BNovO4t3P5Pw6GMMrr/8DsvZJ9jtuXl/sPoMGc2Vh9TMHCw7HIjI6EQYDVoD/4fv1G0SQUUQwUyQmrz8PwCA5fsvYJFTb4y1Y2Yoc3ydrWdsUZtthNJFOR1t9PxfOxw67Q+P1fOhb9WVb3vHVk3VYxiBUAIZ1cUS1YwroEFVQyw49hA9mlRHxDZHdnu1KT5YPkjoKDRBDEaD1iDndx7WjLJFdGI6AMC+XQN2G6F0QXyYCUJRlg+znu00peyHUPqYOHEivLy8sGXZbEx1Vk4SA+LDTCBwids3Rq52xH9ZONK6ZUhClP+yv78/7O3tlXIMgnyQSX8EsXCEC4GgSrS1tZGTozq3HXt7e/j7+6ts/wRCSURfXx8pKSnqNoMgJz4+PmRSfBFCEpcQxELEMkEVFI5trUqxDIAVy3v37lXpcQiEkgQRyyUbUWI5PDy8iC0hEMFMIBCUQs2aNfnWZ86cqRY7XFxc2GVtbW212EAgqBMNDXJrL+1YWVkJlJERNtVCXDKKKTkfH8jVrt3gyXj+4gWGDBkiMv6xrKRc3iRzG/3e85VybIJyyf2hnF4JrWpWUgX7VyeZzy8pbV/lm/VR2r4IBFmI83JSqH2liUdkbpN4Ur7rd0TNgWjfvr1cbUsKixcvhms3yWm7haGuOT2hoaFo2bKl0G1tWrXE1QU9ZN5nRPW+pfK3Ji4ZJQwHBwe52z702wOaptUqlgmln9wf4cVaLEtzDlVoQSbXEAiFkVcsA4Dl17NKtKR4Iq9YBpQ3mV5WhInloKAgUBQll1gGAMvvFxU1q8ShkGCmKOoLRVGvKIp6TlHU04IyI4qiAimK+lDwblhQTlEUtYOiqI8URb2kKKqFMj7A38DC9R7Qseig9ggAS/YHwKDPAhj0WaBWOwjKp2Kd5ihXswm73mNo6Q24r9vcDhVa2CP9GXf4MvLHLzVaRCBIzwq/UJhOOgrTSUf5yguvK4vnkdEwdtwMF48Aley/NMAJrQoATmsOqNES6bGxsRHZ6WHquJFd/hyTDABYdexWUZhVrFFGD3M3mqab8XRhuwIIpmnaAkBwwToA9AJgUfByAbBHCcf+K9i4aDqyP9yXOoWwqlg33g7Jl9yQfMkNG44FqtUWgnJJiwxD1teXCH3JuGxcP3VQzRapjoywAKQ/88c6z+NsWZ1qVZCZTRLvEIo/qwa3RKznSMR6jsSmgJdseaznSBy791Hpx2tWpzISTszD3ul2AtuMHTcr/XglkZTrHuzykWXj1GiJcog9sZBdrm1mAAB4FPFDXeYUG1ThktEPwOGC5cMA+vOU+9AMDwEYUBRVRQXHL7Wou4eZQ1BoBFxH2KrbDIIKaNlEcCJJaWXJpOF86+V1yARBQsnhRngU5ts14Ssb0bGeyo7neSVUoCzhxDyVHY9QvLi8eqS6TVA7igpmGsB1iqJCKYriTE03o2maM74ZDcCsYLkqgO88bX8UlPFBUZQLRVFPKYp6GhcXp6B5JZMlS5bwre854gcdiw44u5frT1y1TW9k5+QWqV2GdsxT57oj19iyTSeDkZiWiYysXGRkFa09BNmgKErktjVbd/OtxycmITziA+au3CC0/tMXr5Vqm7IpfA5xGDBtJQDGLYPDxn0nAQCp6Rmo1H6Qym0jEBTBbDLjerHh4gu2bOvlV1h9NgwZOXnIyMlTaP9hn6L51lMzc2DsuBmDOzZky+q77EJmzu/CTf9Krj8OR8iLD/gek8hXzuumwWH5vvNFZZbcPI/k//2NhzL3AJvFh4VV/6tQKEoGRVFVaZr+SVGUKYBAANMBXKRp2oCnThJN04YURQUA2EDTdEhBeTCAhTRNiwyD8bdFyXjz5g0aNWqkbjMIBEIB2dnZ0NHRUbcZhL8cknykeHPo0CGMGcOfVTEzMxPly5dXk0XKwdfXV6EgBCURlUXJoGn6Z8F7LIBzANoAiOG4WhS8xxZU/wmgOk/zagVlfzUhISHscnEUy2XKlFG3CQQp0NTUVLcJpRIilgnqpGpVZhC2JIrljRs3Sq5USigslgGUeLEsD1++fFG3CSpFbsFMUZQuRVEVOcsAegB4DeAigFEF1UYBuFCwfBGAc0G0jHYAUnhcN/5aOnXqpG4TxJKXp9jwHkH5pKWlCZTl5+erwRICgaAKOL16P3+W3D6lhQu5E8fS09PVaAmhqKhVq5a6TVApivQwmwEIoSjqBYDHAC7RNH0VwAYAthRFfQBgU7AOAJcBRAL4CGAfgCkKHPuvJjY2VnIlBRDn60ooegr/HhUrVlSTJYTCdOrUCZaWluo2g1BK4JzrssTR79atm6rMURoVKlQA8HfeW5YtW6ZuE+Rm2LBhStmPvb09Pn/+rJR9qRO5x9tpmo4E0FRIeQIAayHlNICp8h5PnWS9uYEdh/0wY9RgPAuPQAsrS5wPvIN1u33w5Nx+tl55K2tkhgcDAPSb9oR55Ur48j1KoWPT+YITKyoZGwqUU5plRe6DoiicnNcPq0+F4OoKB+Tl06gz0RPf9k9GeW2mnYnTDjzY6AQLc0PEH5kBTQ0NxPpMh/HI7WLtSw3ahaAnbzBkyR4kXfeAvu00XN46C/FJaej3v+Z8dfVtpyGlIHC7WZ/ZeHJwKWqYGbPb9WxK5N+DxdLSEu/fv0daUjzu3A3BufMX4L51MyoYGOPxg3to1ZIJPa6hVQ5/crPYZQD4k5sFDa1yQuNiKjNBCEVRyPjwGLoWbZAd+RShr96g5T+MK9CX71GoVZ0/KP/12/fRoXUzVOAZXuzjPBWXfHZBu7bwzFHFgRcvXqBZs2ZIfeaPlgMn4/l5L0xdvQMWNati1qhB0G3WBxnPLyH/zx9o8qQRTs/IguPcdfD3XAsAbL3yTXsLPQ7HpSr9wQkAQMUOw5F2/zgqdhiO8LM7cO/5Wzj+21lgGwCk3T+OCu0dVfYdENRHnJcTzKccg0UVfZyZZQOTijpwv/Iatv9URbe1l/B1xzA8/BiL4Ttv4tfuEfiekI5dgW9RVlNDbAbN50fXoW41U76yiyuGI+3WPr6yil0nyGxzwtGZMHHaAQD4um8yzj58D6euVrgT/h3/s2K8KSN+JqLHylPIyP6N+CMz+NpHJ2Wg+exD+OXNZLMrfO84s3o8ACDlEjN5/Ut0IppP2Igkf67rxv3Xkejt6onkADdsOB6IDccDkRzgBgAwsFvALgOAfp+iySabcGSmxDraA9bxPRT4rZuMFfsvIGjHXExspQeKovDLfwvK62gBAPRtp+PJwaWoX90M+rbM5MCUQA/o2RR9JsDEkwuQlpWL7NzfqKSvi+eR0Xj9NQ51KhuivHZZUBSFhBPc7zojOxe6BZ/DaJibqN3ywUnZnRrMRBPOzM6F8+p9CHzMhDGtZFAR17fPw4PXnzBlkw8AICWImYhec8A8fD3HhC/Us56shE8sPyQ1thRkvbkhsU6PUbNx/fA2vrIzV29h5JzVCh2bzv+N+QsWYpMbc1HRKKOFP3m5QpdF/ZYJRyWf8KKQRjBLQ7V+8/DjwmboF6QGTRGS8aikC2aKolghLFf7sqr3l835LBgaShQmjf+H+Nd3RG4vzoLZwcEB3ktGS10/+MEzWLcXnUtJlGDmwBHMkth16gqmDu3FrhPBXDqRlM760cdYtK3HCF/TSUcR68kN2TV+311cePpFaLu0W/uwZM9prJs8BACg180FqTf3CiyPWuWFszdlu3fKc5/IzPnNdroUpvC9gyOUZeFEcCgcrYVfZ4pCMCccmQkT5x2I95mB6uN34/t+7sC4/bozWObQAW0smOi4xk7M500NkpzN70boO3Rv2UCgXF2CWRw3X37BGPcL0Cqjifd7+e2TVjBz4AhmeSkKwUxSYxcBhcWyMtmyldk3RyBrlNFit/3580dlx1UmPy4wT4gpgTuFimVC8UOcWC5tiBPLyoRXLBP+XjhiGQCfWJYGD99A3Hv5AXrdXFC/RmXodXOR3EhFiBLLykKUWC5K4n1mwMR5B7Jy82A22gM0DZg478CDiJ+sWJYVYWK5uNKtSS18OThTQCz/jRDBLCMhT1/yrZe3skb4+88obyXghaI0OL3Iwt41NGT7CUdsZYZGgl9+BQDk5fMLbmf3SwrZCgDbTl5Hq7Fr2HVOrzLn/W+hcrUa7LK2rh7rfqGhVY5dLg68jvgInTqtUKdDb4S+egPHKUyPg06dVpi1wg3vI7+q2ULVsM7zGN964L1Q6Dbro/B+W49ger7ajGC+R44LRsUOwzFvq7fC+yeUPGiam7radNJRdF7lL3cq69Sbe9GxiQVSb+7F08Or2V5lzruycTv7CACw3u8BAGC7/1OYjfIQ10QmsnPzYGDHnCsGdgtgMXI1u15ciPeZgXifGYjxng6K4q7LwnqfyyqyTnl0mM9keO04/xAAwNhxE9/7g3fKzfbXdhyjE9qNZ971babwvRc3iGCWkU6tmuBBGJO0gaZpZIYHw6p+bWSGB+PLj2gJrVWHtK4103q3wJFb4YhOygAAlNHUwKWnn9jtPrOUIBga1UYzi+rQt50G25lbBHqU/xbhbNO9O/YfPISO/+uKnIxU/MnNwrkLFzF18iSFXDeUzeUbd7F0pgsi719Gy38a4cRuZpgtO/Ip3FctQP06NdVsofKhaRpLJo3gK7Pt2BJlFQyjGBL2Fjm5eajYYTi+RccDAOu3nHb/ODbPGa3Q/gnFG1HRapx230Ss50gcvBWBWM+RuLvCHrGeI5GQlq0SOwaOm6WU/dx/9xO+997BxGkHPK89BwDMtG+F/D/M/Yb33iEvVqPXITnADR2nbUNygBs+HF3O56tcGqBpGoucuW5dPWcxo8b9Fhav0daIHwl4HhmNe5uYMHkJJ+bjcPAL1oe5fYNqSjtWyMsPyPmdB32bKfhWkPQlJWg39G2mICVoN/ot2KG0YykL4sMsBdGPLkK/YgWZ25Vr1F0F1shGVFQUtG/IFw/TZtNNhL14KbaOtD7M0lDSfZgBgP4t3w3w4dMwtG/fXsnWCCKLD7MkirMPMwBkvlBej44kH+Y2rVvhxo65Mu+X+DCXTCZOnAgvLy+R28/N6YFOlmYit4uj0sQj8pqlEJmZmcg6u0gp+6o23hNZ2Tl8ZfL4MIujqCb9yYo0PszCMLObh6xs1TxAiUOSD7M4/jYfZiKYSyGJiYkwMjJStxkEQrEkKSkJhoaG6jaDUAIIDw+HlZWVus0odmhqapLY7wri7e2N0aNHq9sMPoyMjJCYmCi5YimGTPr7yyBimaBMxo0bp24TlAoRy4TC2NraCi0nYlk4RCwrTnETy1ZWVn+9WJYEEcw8/I0Xx6JIZfk3BqsvTRw4cABly6p2NnxRkZubK1W9ovq8enp6RXIcAgMnHmxhAgMDi9iS4k1xT47UqxeJNqNswsPDZapf1BG6isO1slS5ZEgTL1kc8vgcL126FGvXrpW5XazHEJnbcHjZcCJsbGwk1ku5pJyJE5dTq8PRUbSv5ZgxY7Br5iCFjlG+mR27fHfPQjStV12h/XHQU+EEQ94EA3S+clKIU5qKTTrjUKZMGWR+eSFTG61q3AfGnE+PFTq+dt02CrWXley3t+Ruq9OwKwAg83mAXO2fv/2EDo7yxzrX0NAoMeEhlUXsrqEKtTedekrs9rp16+LTJ8UnpMlL3P6xCrWvNP6gkizhknBUvomAxiPdhZaLS7AiCmXdk/T7LJDbBnkop6ODGP/NStmXKu9JABC3d7RS9tNs8Tn8jE8BACSeXqyUfY5wO4MrTyLkbl8Uv/df4ZLh6+uL8lbW7AsAKjZhhtlEfcGcepw2kgR3fkq0wGvV/GlIifqE1UtdoaOjjfyUaNSvV1fsfmI9hsBshh9mHnsKsxl+OPv0G7vNbIYfzGb48dWvPNOPr1wasczbq2tgtxAGdgtx6/kHdJ7BDSZf3WEFXxsDu4UCbYz7LkJvve9ij7Vr5iDoNreHbnN7AIB+q/5oPmAyW8YpL4yo8qb1qiPoyRvo92AyIDmvOQD9HtNx4e5z6PeYzpZHRsUBACZsOMyWX7jLzOQ2+ld+AQMA+SkxAi+KopCfEoNaNaojI+Yb8pL5o6K4b2dm9Q4e4gCNMmWhUYbppeS8A0CXrtw0thplysK8GvNgUE5X9kml4sj88gJdB4yEdvXG7Gvhms3Qrt6Yr5529cbo0GcotKs3Ru4Ppoch59Nj6NRry9bRqdeW7+UfdIdve+E6fpeDlPpZpKVco27wv3EPAPDsdQRbxsu567eFlnPg/b8WfhdFs4bc8z39/jGpXhwyHvsh7aEvMh778b1MDPXw7fohUBSF6JtHELx/nSxfRbEmdtdQmE3zhdk0Jv0zZ5n3VW/eOb42F55952vDYft24YmV1C2WTSccwtXnzHXddMIhgVfdGfwh5S48/Yxa047AzOUQfiVnKt0mBwcHDNxwFiZO2/lenDIAsFvL3F8465w6HKFdxcwUsR5D2FfMjsF868YVdNhlYaRccoOB3ULk5uWz9xrjvovYew0AdjtnuTBmA5bwrWeEBSDz+SWRL9uOLZH5XPFQqRyxzHv/0e8xHQY9Z/CV8cIp591u0HMGe2/OeHRK6Gv5pKGgKApRQQdRUbc8Mh6JfzjkxcHBAYO2XQcAmE48zL5yfnNdaPq4XeFrYzrxMN+y6cTDuBH+E8//G8CWGzusBwB0XXAQxg7r2XUA8Lz0BL2W8U9UNXZYj+xcphMp+HkkW//YAm7HWsqVLWJfLw4tRsqVLWz9tJt7kXrDC2k39/K9TAwq4sv5raAoCmk39yJop+D/RlmUmh5mX19f2Dc24Ssza2OP11ePIPJbFLqNmM6mrRaHuF7m/BTpw8Zp6lcWua3wBSUzNw/33sehY/1KqD3vPGJ2DBa7b9PppyUen6IoJAcIRse4eO8VWlnWwJfoxIIUpEyd5PQsGFQQHRuY80QvDGE9c1U6D8PLC16oZc0E5c8IY4ZCOw6fhXvHBXsseHuYU4UkNgl68gaDluyB1wInDLNpA4uhi/Hh1H/shUhHqyxiArYKtJPnaX7IgL446b1PckUAmvrMTHhhPcyGxiaIjf4F7XJMWuk/eUw681at2+DpE+E9uMrqYeaIXwBITE7B6BmuuOizB9rVG+PlTX9Y1qsNAChXswmyvnIjoWhVsxLauxyfmATLLgOQ8OoWK5azPz5Cn9HTcclbMCarOnuYyzXqhqw3N9G6/zg8OX+Ar165Rt1wcvtq9LPtzJYJ62FeuHk/5o4ZjOfvPmHwjNVIDb0A3eb2yAjzR9Uujvh5mz+zH+f/yyuGxVGhAxPSLuOxn4SaXHTbiL8ulBSE9S4nZ+bCoDyTkMlsmi9ClvWChVlFmE3zRcxOB4H6knqY1Ymw3uWkjBxM2X8HJ2bawnTCIdxbMxAWlfVx4MZbLDn5CNGFegWV3cPs4OCAPX25ae+rjPbAtwNTYbP8BMK/xSO+IP2zidN2dpkX45HuMo2KCrtHiepdPnL9CRbuvYgovzUwsFuIbs0tcG4Nk0bbwG6h0PsY534krRgur2B8dWH3JACIT0nHJLcj8Fs3mb0XpVxnrof6Paazy7zo95gOmqZlEsK6baUbkXFwcMAum/IC5QHPvqJF7Ur4Gp+GfpuvItZrFADmvCujQaGCjnAXtEou3gCE9zB/j0vBLK8r2D+rH+qNdUeC7yJ0W3gINzeOQcCjCNi1tRS6T6Mh/wEAnxgWh34vJvpQmgxxxisqkMznr+hhFkbMY39UMjJA22aNBMSyvIlG3rx7j0atOgIA7j18jLj4BJjVaaSQneW1ysC2cRWU1xIUS8J6nBWhb8d/YG6ijw6Na/NdiMSJZXn4dfckKhnpIyPMnxXLAHDvuDv2+10R01I4Nq0bIeW6B4bZMELsw6mCk+66B1KuewgVy8pk9YbN6NTDTnJFHpIS4lG2bFn8yfvNimUAePrkMX7//i2mpXIxMtDHRR9uOB+OWAak90MzMTJEwqtbAIB757yR/ZFJZnDJ20Ogt1ndZL25CQACYpmzjVcsi2LjvPEwNTZAj44tkRp6AQD3oS85NV1sr7Npd0YwtRnpCgAYsVh8ennr8Utw7f4zVGg7BBlZOfjfaFdU+t8IsW1KGxyxDAAxOx1gYVaRXV7v/0pdZikNQ11tnJjJjHjG7hsDi8r6AIBx3RsKiOWi4Jf3dJTV1MDtdSP4BLIwsVyYyjOZ+xHnvvQ+OlUhW5x6tEaUH5O4IjlgIyuWOeuSMGk3EIH3Q9FqMJPsYvi8/9htT1+/V8g2icfWrwC/dUyoM869iEPCFVJRYtkAACAASURBVPHnPYfUdGZEYfJaTxy9dJtdVyZ2LWrC3LA82lvwhzo0KK+FOjOP42LoF5n3Wb2SPs4sHQZDHu1wcyMTv3n01rOo7Cj5t6vcnwlj2G4y04PvtI7b211zyDKB+gf87+D6o9fQ6z4Rmdk56DL5P5j1Krq8DsrpziqBSNPbLIxGDepj2OABoGkaHdu1QdX6TRAT+UZpdhXuXZbU2ywvQaERsGkp/AlQmXB65XgZP7hkTNgoY1AZecnRKGPAjBYUdsFQBHVNosv5/lrsujS0/Kch3zpHPP8tFP4/FyYrJxcVOzIjKwadnZF81wcVO45E2r2jGL7YHcf/4/clfRr+EYNmr0f6I6ZX7tnbT+xy93FL0KxBHWydX7oilcjCIvt/1G1CqUVUj7I4zs3oCoC5N5nN8FPZPUpaft45BaO2A5ARFgDd5vydGg3q1BDRSvWU0ZSuP1KvQnlUaDcM6Q9PokK7YZi0Zg/SH55UmV2c3mVR6/KQ4Msfvzv+lHTxvLNzf8Og9zwAgLHdAiQEuMGg9zxUNtJD56aCrq29OzSB5ZCFSL3BxEAPi/jKLgNASnom9CsI9rAri1LTw9yqldAedAF/Zs76o+eyi9z+jswfa7nrPNYP6ed7wcQe4twxxNHX/RYAoPESZmjYbIYfFp0Ow+e4dLn2J4z6I9fg6uO3aNOAm73N6T+u/5EwvzFZGbN4M6p3HY5hc9YhI8yfrzcuLz9fok+oMMZvOIzag10xYiXjKlHYZ0yYD5m8nD53EQBXIOclR8stlgMu8Q8ZDh7CDC9//Fg0PpZ12lgj8ivXB72wD7M05BdKn87bo+x+QDoXhKLk41cmfat/cAgAxg1DlM+yOHj/p5zlhORU9Bwn/maQdu8o+0q+68OWARAQy/odhiLlwSlWIAPgW75xYB22zh9XatwxhNFiWQC+xHOvcbx+yoV9lksizRb44ktcGrtuOuFQkdvQsqVgkiGOWDYbtYNdr+S8HaO3X8J/BWmweeG4WbSvy3V9FCaWX35PltquRqP/41vnvf/YL5Y8BD9k1hpoa5VFRhhzz8wIC2CXAaBCeR3U7zVGantkwWrkcnyOimfXee9B+j2mIyFF/H2b182CI5DTH54UEMvTPKUfkRX2OwNA51XMKFnN6YLXa7NJhwXKpOEPTcNiHNe9ktevWVqSL29mXwkF2R2TL2/Gu6PLcXTpaL66hraTUcXYgE8g8y4DgH6F8gq5Y0iEpuli+2rZsiUtCxb16tIA5Hrp6GjLdCxFkdfOpUuXSn0MvYoV5D4O70uVn0fU/pVhtyzflTIobvYq8jso8jkGDRqk1M+hqs8r7L+nqvODwI8i/69Hjx6p23yJFKfrAIcqZqZq/39ramoqfJ10cXFRqk3SoqjdAOiKFXRVbmcVs0pKsZWXQQMHKn2fxRUAT2kRmrTUTPrjQNM0iftbjAkNDRX5FExQDjNnzhQZPaC0kp+fD01NTaHbXr16hX/+UWxY39fXFw4OgpPPCARC0aCrq4uMjAx1m0Eo5fxVk/6IWBYkKipKYp2LFy8q9ZirVq0SWk7EsupRtVguTudYUBATwk6UWAagsFgmEBTBzk62ycJFQbdusrsoqRsilmVj5kzFQqsWBcW5w1YYpUIwa2iUjI8hrdBQZk9W2bJlYW5uLrFe3759lXbM2NhYrFixQmn7E8aMGTOkrlumTOmf26qlpSW5UgH169dX6FjixCkvXbp0Ueg40iBNTPI6deoofJwDBwQjbgijOD1MENQPRVEICJAuIU5qqmLRJqTl69evuHnzZpEcS1UUZZQhRVDn9UBSx4m013FpkSddekm7XhZ7l4zQ0FCxdYYNG4oTJxgn+aw3N5CQlAJjQyZkT3kra4nRMBxnrsSccUPRuklDmTL9xeyQL7Od162PmNi1HsxmnGHLpPnTcH6nlKDdsJvrjpCXH5AStBsPXn9C+8bc2aQbfC7B1Zkbc1LfZgpSgnYDAPSsmfA30mR8slvrh4ClzISO0/feYUjHBgV2ACZO7jLbDQCW9evj/YcPYut36fI/3LrFJJdIDdrFlq/3uYxFzr0lHi8xNQNGerrQs5kKAEg6twoJqZkw1mNmzhoNXInEsyvF7mOU2ykcXsBMyDAcwAj/ixcvol+/fhKPz/t5JX0/vHVzvwtOHv3X0QVXT0gfexIAtGs0Rc63F9jq6Q3XdVulsqOwLYXjMK/c6omVcyZJbYNRk67IyMiU69jS1Hd3d2d7T7Lf3RZap1zDrsiSkP2Pt85J/0CMns/N2CmL3f369sWJ1RMFtus2t+ObgCSKql2GISklTWI9VRK9STByzR+ahoaE7+FmRDyaVddHwxWyJauJ3T0ct95Go2tDZoK02dQTiNnliODwXxi++xZidnEzi3K2cTCdclymYymTuH3CJ5D5h36Bfcta7Prn2FTUNuVP5Wvq4o1YnhBylXgm/8lyrRBH/OGpAmWfY1JQ20yfryzo5VfYNKkJALjy7DN6tagNk1HM9bbLPzVxbgl/p82Adb5smd2qkwhYMYxvu9GwTVLZBwDJF7kT/YKfvYd1C9ke4A36MjGBu3b5H27fuSu27ty5c7F5s/wZ+mS9fqVck290T7+n9L3BstoUu3u4zPbYbwmC/1wbgXMt2X+DVO158zpUdViBn76rYGDvKrJ+qPsE1K5sKHzbx19oWa+KQPmobRdweDZzT+698jgur2Q+p9EwxbJJinPJKNaCublFdfrW9pkw7eeKvPw/SCyYRWm/yBN3X36CjlZZRJ/7D/oFYUmKMjW2vIKZA0cwJxzlBIzfAYoC4nxE95waj9yO1OA9IrdLQs96MiiKkjmMkKAd7sh4ehYAUKE193tIf8J9COg9eQUu72HcMnRbDQQA5P4MR+jLcHToPRRD+/eGz07+P/aQ8TNxev92aFVlUjTzCmZZ4RXMisARzABz8d26lRGhhUUeZ11PTw+pqamgc7Ogoc0Nb/Mnhxtb07pHLwRfZ2Y+U1rMRUWYYFYUrepNkJf4A9eCb2H+stVISEjCz4gwdvvQ0RNxyps7y7iMUTUAiqfFBpjEJdkf7qNc/Y6oVsUUP37FIuv9PXb7w2ev0K4F11VCx6IDACDr1TUAQPkm/yLz5VUAwPA5a3F861IA3DkK5f7pCUC0YJYHnQZMjzjnvw0ANWxG41uQNwDg0csItG3CH4pRt9VAHHFzxaAekmM7i0O3uZ1ahyeFCWZZqDxfttjq8tzEORRHwSwPHMGcdGY5AOBxxA/0XuqN+NPMf91o8Bok+nFj0RoOWi1xn8IEs7SYjNoFX19f2Px5InNbeQWzPHAEMydCgr71JKQEewIAnFZ54cgK/odXve6CD7PSkHZb8qiSie0kxAd6omIXJuyjqgVz4qmFCH7xGcuO3EBCahYi9nJjEBsPc0PCyQU4cuMlnLo3gdFQJg6yMs81aQWzMMQJ5sSTohOjyYoqBXOJ8GWIvbCBFcsA4L9+EpIvbUL0OfEn3psPn1HeyhqBIdwLgGHzf1VmZ+WZZ1F55lm+siqzzoqozU/8kRlixbIw9G2moOGwxZiyyYevzKjHNOjbTIG+zRSJ+wh++RX/zDgA3nu1idN2BL/8yi6LI/3JGfbFy+U9q2DccZhA/ZZNrJDz47WAWAaAMcMGij2Wvu009sXBtLfk3nJe3n6LhdHAlQgO+8iWVRnK9Cwu874utu2WLVvY2bK80DSNf/9l/ldpadyewj85meyLl+DrV9DLTnoXGO0aTaFdoylfmU7NZlK372ndFS/v3+ATywBYsVzWuLrY9q7rd7Bpr3mRJgV21vt7+HD7HJ9YBsCK5dR04X6JHLH8LPw9K5YBJm6nPJQryOZXoXF3pGVInxiAI5YBCIhlcbz5+BUdHGci8D53hEyvJfObj10if49XUVJlwVVUWXBVoGzqCeU/3BXGfDozahj+MxlmU09IqK1+Oq88z7duWpAhzdTFm12WRBvLaog/vRRhn34BACuWc34LZhGVlkqjd6PS6N0Sy0Rh7LgZxo6bJZYpSocZO4SWG/ZbIrScF45YBsAnlj/9jFXcMAlY1hDs+eTw9ssvGPw7C0FP37JlZvbzFD6mddPauL95HJ9YBoCEAtHZv30Dse3Npp4QOKfMpp5ARo7s/zPDvotg2HeRxDJ5ufL0I4wduQ9j5x9GwNhxE4wdN+HVV9X/voUpEYJZXhpZ1EZmeDCc5jFZhMpbWSMpjLkBXLn9UOnHi94+EMcndcTmK2/RYS0jvn65D8SNtzFKPxbAuGe0saqD3fOd0X2aG+zmuSMlaDcSr+9EStBu1hVDHNZNaoIGDYpi3C36rz+D+CMzYV0wXPdl32SpbOHtaeaQcE+24Ou9bST7vKYE7sQLn5Xseuxlxj1EX8oU2A1rmCLx7EqM28JkqTIauBK/TjGCrFPjWjLZy8vVq1clV+LhSoD0kyyrVjFDzrcXqFCHO2Ey++tzmY7HQZg4/p3wXUhNLhsWzUD2x0fI/vgITXtyY4cO7m2DwROlvwHs9BZMBatXQVdsmxZW/MO15bSl89UOC2cyfHGEctprxjUr/fUN1O8uXZpZRWhUryZevPuE1buO4tmbD1i2wxupocxvfuryLZUfX1n8cvsXXxMz+dZ3OTZR6jHarwqA2dQT8Ax+x5ZFeTAP21ZVDfjcMYoraVnMg1zVyUznBcf1InbvaFhWMZBpX83r8osw7bLyzcHIK4ifHufN33ES5z1FoExc+4QT/Od4wol5AmXy0nfpAVQauAxvvzL3yJgkftekpAvr5N533aqmCtlWmKDHgkme7h0QPVdnwBJPJF91h3XLBrBfyIyWxvgr90HD87JgFLGK5SRfI2N2OeIrT+zzmF2O0NWW7X/G+X8kXVyPL9GJbHnSxfVIuih7TGZhaJXVhF55bQCAseMm9G9niYQT8/Fwyzjk5UmXpVaZFHvBbNBnPgz6zMe3mCQ4rfPh25aWmS2x/bPwCEQ/ZG5UmeHBbAITm47CE50oSveGZpjXqyHuL+3B9jgP97wnuSGAE3e5T6KpmbkAgIS0LNx98wNhkcJF9+FlTBrRGzsXIGBzoaQIUvQwA8DrHcw+KAo4v2gQbrz6ym6rNUG0C4hJx2GoYTMaFVoPQsDulQLbh80TTI2ZkpaOGs27Ii4hiS2r2oQZztauJj6pRkrgTsamKtzA+efvhEktlnn5cpQZHko8uxIX7ofDaOBKOG8UFHS87Ny5U2ASD0VRiI/nBq9funSpQLvzF7mZ4WrVkz27YuQj5uErPZLpreT0OGvXaMonokVR0bweKls0QVnj6rh2TrHeuhfXmO8o8C7zwOnnJf4G0GMk89uUq98R00YzQjWXZ8LO05eCCYRS0zNw58lL5BbqWdt97ALmbZTOJal5gdDm+CiX0dTE2au3UK5hVySnSucvXPghcJG7NwAg9M1HqXq6M8ICcPfYNrRoZIE1M0YDACq1HySVX3Nx4JcbM2pS04hxLeL0OAvreVaEByvsELPLEZOsxfeMFWeebxwCAPi5x1lg243l0o0mbfS9jeM3X7DrCWmZsHaVbrKpMMpoamB+/9Z8ZZze5Uqjd6OGi/j5EWU0NbBgUAe+MmPHzfid/0dpvcwX145D3Nk1rDA2M6zIbpOmdzk1Iwv1hyxAbh73WpGRlYNRq/cpbNuzd1/gd+Mx9LqOx+LdvvAOuMO3PS0jS2z7d8cYd0CKouC/kesqY/DvLBy8JJ0mEEbf1cw13HiYGyb1ZnRMbp70E+44D6A1TSoA4PY4m009geozxd8DeSmjqcEK41qVjdhyo36LldbD7LDBD58PMCPvCSfms73N7eYegM3SI+KaqoRi7cMsaxzmkujDLCuamppIur5T5na/8/Jh3JMRL9JM+hOF6SgPuWbDykNx82FWJcr2YV6x+wTWr5fvKV9RH2btum0Uai8LqvBhlofM55ckVxJD+WZ9JFdSIUXtw7zOoRUmdJUvWktp82FWJor6MANA4sn5MrdVhw9zcaQoJv3JCvFhlo0SO+lPnsQlHMqWLVtiQs/IQ3Z2NnR0dNRtBuEvxsnJCUeOFP1TfnGmffv2ePBAMK0wQX7+/PlTYkKHFid+/vyJqlWrqtuMvxZLS0tERESo2wyCjCg06Y+iqIMURcVSFPWap8yIoqhAiqI+FLwbFpRTFEXtoCjqI0VRLymKasHTZlRB/Q8URY1SxgcTR1GJ5U+fPhXJcQojSizXq1dPJcezsLBQyX4VQdHA7KLC8wwePFih/f4tELEsCBHLykdDQwNVqoieXEUQjjLF8ps3gq5TAPMwQxCkc+fOYsWyvr6+yG3Kpnnz5irb96RJwsONGhoKDxFX0pHmsd0bQOHQEq4AgmmatgAQXLAOAL0AWBS8XADsARiBDWAFgLYA2gBYwRHZJZlOnTqhbt26Yuts2iT9UJUy+PHjh0r2+0FC/GR1oGhGO1GjK35+fpg+fbpC+y4q4uLiVH6MY8eOqfwYJZm9e2WLlU2QnV+/fmHNmjXqNkMkFStWlFwJQI0aNVRsiWpo1KiR0HINDQ1cvy4+upAs5OXJHxGkuDB06FDcvSs+PnRKSkoRWQOEhYVJriQnnp6eQsuTkpKEliuCtElOHj16pPRjszZI45JBUVQtAAE0TTcuWI8A0JWm6V8URVUBcIumaUuKorwKlk/w1uO8aJqeWFDOV08Usrhk8MbClZfoLfKnMG238S6+xKruJDh48CDGjh3LrksTI1IUnJiRwpDme7SyskJ4eLjUx0tPT8efh4elrl8Yjj+ysrG1tUVgYKBK9i2K3G/yRbfgoFVD+nByRYUyzr2iQNl2Ft5f5ovLcu2nfFPJCXkK82ujrVzHKkyVhdL9/8+dO4cBAwZIVdd7Qgf0biJ776bp9NPs8jeP4SinpXiGzsvPv2HUnlsK7yfh2GyZ6rea54PPvxIAAPPsmmFhP/l7+ZTp66yqc1WW/SYcnyv3cU6nWrC9mimXZO+MajZhEz5HSR+OLPr8Bqkj80hCv9cchdrHeck/MF9povz3X1FI+s0nTJiAffskT75saFkf99YoNieMFyMHZu5O586dJT64CEOcS4a8VyQzmqZ/FSxHAzArWK4KgDdG1Y+CMlHlUiNsQp/d+PkI2L8J95+9hlbZMsh6c4OduJf9PkSq/erU78R8iC12qDLvEn5tZibfjPEOxcfYdNxdwD8RaMT+x7jxLo6tx+HhQm7igiSeYPN89q7wgXbZMgj7GIVI73kwHMz0mCRfEN1zUm3YWhyYNwQDW1miYgVdpKVn8InlegPmIDv3N35c8sAh/zsY2asjypbhprzU6zoeqbf2w3nFHqyYMBDNRy5B6i1GNJfR1ESMN78YjfeZjgrltPB17ySYj92N7k1q4OgsOxg77WA/W8iqgUI/Y2bOb7gevIYdk+3YzwYAfx4exuEr9zGqVwd8jopHbXMmyoW+7TSsnzwIUwZ2AwAY9ZyBxGs72KgXnZvVR8AmbmzqjEe+Ir+nnpNX4tqelQAA6h87lC9fnm974mnBGdenXNqw5TTNRAkxGsLM1k65uk3ksXrP34XLm5jvTf9f6W+kFEWhdbPGCLlwBDq1WiD7yzOBOto1mRtrztcwvuXC/P71TqCMg5FFSyR+YCJqdOg7Ek+eSH7ozIkUn6zgxIWrcOz3L7TrMLPueScGZn98JHKioDInAEo6p3ce9sW0UQ7sOZ397hbf9qy3N9mySi17Iy6UEbgROYZo2pQ/znWmmEgWZp2G4Pimxch45o8NGzbA1dVVqFiu33MU3l/jv1HVth6Bz8HHoNusDzKeXwJN08h8cVkm0cwrlgNexcLuHyZ8lsP+Z/Adz3rBwdyViZMdtcEGQe/iMerwc/xcb4MzYb8wqHkVmLsGYUioA3x9+c+r2F2C8dM7AojddQr93W/g/CzmGms6VXjYSI5YbrwkAK/X2cFshh9idjBuTlm5+ag17xy7LopyWmXgFfwW5bTKYN6xh4j1dBJZ13TSEXY7Z5nz3ruZ4j26Dg4O2NOvKkxGuiO+YMJ08MsvWHH8LkI2OMFkJBPaMp5nMvXTzdxoGQv7NceTT7Gwd7uCaB7Rs8z3MdY4tIHl7BOI2OaITivO4f2vFByfbgObf6ohMT0bPf4LwFNwRXPCMcEJ2w8jotDO0hwA0HzWQYS5j4XxiAKbDvEn7Yg76IL4QxNRa/IhnJrTC4M2X8IPr3EwGcPEZH9/YBZM9PivnQBw88VnjNl6Dv/UMoX/qpGYt/8aDl5jrjEJx+cg/thsJBznCsKHET/RzrIqvsenoroJk/HQePhWdnvwiy+wblqLfQeYzJI3X36FddNaMBnB1I0/xt3n2O0BODhTdKfWhuOBcB3OnBtDVh5E4NN3SA5wQ0JqBoz1mPCVz/fJNqGxnLYWDHrPhfv0IahsrId/2wjvaf8UFY+W49ejZmUjvDi4BPaue+C/QXhI1tRg0eFee8/ZhstbmXuKnjU3ylWc1yiYTvJBrKczKk8+gug9Tui18Qqm97RC72Y1kJXL9MzXnHEcsZ7O8HsUicFt6wjsP3qT6DwUAzwf49wk5npdeT4TBUdfryK+HlkoULfasDVIOr8G2mXLIDcvH0nn18CgHxMhinPv3DyokdD7KCeDKOfeySuWLz58h92XnuDqGv7z/c7rL9AqUwbtGjDJtYyHbsC6UdaY1Ju5HzVw8cC7vcyocKLvIhg5rMfdu3cRt1905yAAmE44iNh9Y1FpvHQdkArPpKCZRwylPbJSFOVCUdRTiqKecoabRUW/CNjPPGF2aNEYyc+vKXxsXhHs3K4GIuMy2IQeTVYxN58b7+Jg28gUbf6TPSJHwCpnnFk6HJHe0sew/HFyKXq2YkKRfT8mOMv047mtsG3LhGObucWHTywDwDs/5jvyWTUZdauZIfXWfnZbYbHM4ete5gk+6uAUHJ3FvUAVvrEWprx2WeyYLPyCNqoXE56otrkJ9G2n4dWnH0i+7oEpA7vBwoEJQZP/5w/+0DRSAnciJXAnn1iWBEcsAwD9SvawXbKktOeIZUC8sBZGyAXG75cjlrVrNodu3Vbs8oBe1qxA9tu3TahYlgRHLAPA/YtHZW4vDMd+8iX8UUbWQGmZNspBcqUCOGIZACy1ZRs+jAk5Dev2zMPMjJ7CQyEudT+EnzFMuEFdnggYn4MZ95aMgqga0g4zioIjlgEg5CM3Fqq5axAuTm6NqA02AIDulsb4uZ5ZHtSc8QfmbJMFjlgGhAtrXl6vs4PXTcaVa+7JUJjN8EM5LU3E7BiMgR63MXKv+NBaE60bwrmzBSuGzacw/2XTSVzf+S9xaYj1dMLHmFRM2HeHrStOYMtL/NFZMBnpDpOR7hjqdh7vfjA9yJ0aVUf80VkY6nZeZNvWdU1ZscxJZuIV9AamLt6I2OaIgVuuImTVAMTuHQ2bfxhRYFRBB0//4z5Y9G8nPKoIRywDQJj7WKF1CvNlzxi0taiMH16MoOD8D4WJZQDo1rQ2vhyeA/9VIwEAm8f3LLBJeJjMdpbMQxNHLBeGI5I54riyszs0KApD3c5izHZ/xB+bg/hjc1B/IldcihPLBnYLsPlUMLvMEcsAWLEsL8mXt2B0r3asWB67gfn/GfTm9pTXNTdB8uUt+BqdiMaj1ooUy5LgiGVAUFjHejIPYdF7nHAj/Cc2DW+LXk2ZB8JWS8+yIzJ/aFqoWJYERywDXGEtTCwDwI+TTIdZjN9KJJ2XzVVKQ8w1r2+7BqxYNhvuhtjkDBgP3YDB606hXYNqMB7KROloXNOUFcsAWLHMiyS9AgCx+6Q7X1jbZarNJabAFQMF75wxjp8AeDMjVCsoE1UuAE3Te2mabkXTdKtKlSrxbStvZc33HvJUfCiuFn2YL76lHfNHK2fZme9dHN0aVMLPTb1ZIfVyBXNz+bW5D3zGtsbjxeJD0LWfzfj2dJjDPLkbDVnLvsckpQvUz8v/A5qm8fYb81Ua9l+OdtM9YNh/uURbDy1nehF4xTAHcxPJruK8riSWU/djwu5rMHH2AAB4XBLsBeV8DgBsVioAsBy/DdWdJId0SQnciX/qVmMv0h9817Pl4k4mALgX9pZvvUK7oQj/9A0V2klORvHmWyyMHf5D40keeP7pF0ZvYbIwGjsID3O0zTcYdYYyF4Z2k9xgUDCkZqDg0BovOV/DkPHpKbt80pMb29S+R1exbfPy8kHTNMLfMaJEy7whmna1g5Z5Q7lsad5zKK7feYAW/zJCSKegd1inbhu06i0YmijkSRia9OCKVE4mwMIZAZXJ6/eRuH73kdTndF5BCMTwD5+Zeg274fX7SJRr2E3isUKe8Scq0G1hj/CPX6Dbwl5i27WzxrCiOEPBkHPSwiuAozbYoFVN7sQiSedVYXZcf4uOa7gPFWbTTvG9S8vEbhaI2TEYW4a15OtVPju9C466dJRpX1G7GbHGK4ZrVWJ8h+uZ6WHfhP/JtD95iD86i+8FAOcXMz1kpxb0l2ofsXtH870A4Oxc2R5IPa8yD9ImI/nncTyMELytNpzFfcAwn7AflcYyPveVxu5llwuT8zsPxgVD28YO62E5YTu7Lgm3M/wTX01GiO9UiD82B9E+s9jlQzO559d7L+lyCSQHuCH+wgZ2OTlAsdBi4jjoyvz/ki9vEbTj8ha8PiwYi78w915+RKsx3FTn+gVuh/pSuh92t6qKxtWNWH0S7sZch2M9nSWe6w8jk9BpE3e0rsqCa3zvwthwkukgXHeceShxP3MHJgOZkKuG/YWPqHPuk2EfuM4F9YdLH6Y15vgCmBroIuGUK2JPMMI94RTTaXjbTTahazrhIADg+RduzgSruSfYcmmRVzBfBMAZWxoF4AJPuXNBtIx2AFIKXDeuAehBUZRhwWS/HgVlMpEZHsz33qmV6IxTdx8/R07ub5Sz7IxvP6MBAFkRd1HOsjOyIu6izxjph9L33f0sk5333nxF7u98GA1Zi+9xjBhN1CFEowAAIABJREFUPL0URkPWIvH0UjR0cRdo43EuBEYDVqBhDabHKOn8ajz0mA6v2crz7RFFLVN9ViBH7BqPcw/fI95nOlb73kf/tvzRMRYdugaLsVuReHopOs/by5eVKmL/bHw/orx4isLo2Lwh/tvP9XVMf3gKVnVrIP3hKRh2Eh1vctGhQFi7HkKC72IkpmWhWd0q8J7LpOJO8BUe13O2gzUiT60BTdN46LkAyVeYYcJ2jWoDADYcVXxUQxG27DkA7aqNYNWA+Y1yo97ixa0A5Ea9xeu372Xal07dNsjJ/Y30jEz2fMn+9BhTlvyH7E+P8Trio0CbTq2b4+V17lN89sdH7PuslZtw/Y5yI0aUs+yMjgPHIyMzi++cnrrMjT23C7PtwEms3ekNKwvmN8t6exON69dB1tubzD7FCOdOLRpjnRc3DmnGM39Y1auFjGdMIpquzsrJdlYcmdGjIe4t682OsMXsZB5IR3ViJjn32iw5LTovZjP8lGrf38zwLRcx6d/mOBD4AvFHuZGCElKz2J5dXro0qoqjd96h97oLiNo3HnEHXXDp2RfEHXSBVXVjfIvnT+Rj7LAeb7/FIcGXGflL8F2ExUO7IMF3Eao7CU9WMnwzIwEOBL7AgkHt+bbFy+j//TfQsUk9PD3E7QxLKcg5UF5HOf7S4mhXxxAh8zux67/cekps4zqsO3wCQ7FkONNZmZKRjfizTI6Dwj3Mrp7nUHfoMiRf2YqOUzajuQW3r/T9ccXyIsjKkpMP0WD2ccTuG4uuq86jWS1u0rPkjBzl9zBTFHUCwAMAlhRF/aAoahyADQBsKYr6AMCmYB0ALgOIBPARwD4AUwCApulEAGsAPCl4rS4okwpZEopw6NymGcIDTyIr4i7iwrizeLMiGCfwS4fEP/W2X8/cUKvMu4QJnZmb7XifUHFNWDo2qonQnVOReHopn4BMPL2U752X2YP/h6TzqwXKHbpwfSu7LT8tsB0ALAYyQ0MBIdzhe72uTPY+p+WCmdE4jvo1JnId8uN9pgssL3fogOom3NnfDg4OWD+mJz4cZJ4c7252EWqPJCr1Yi7y8mToA4DF44cILU8KYcSNblvBofn1Y2zx6zjzlBp1TDZRX3jY/OoW5vtxHdlTJh9mUazZJnym8bVb4oesF053QW7UW6HbGjdkhm9X7pQu9Fv2p8d4c/MsBvayRvyrW2z57nWL2e2y4L5yPnr8r71SfZizIu4i5fUNDOjZle+c3rVmAbu9MPNdRmDptNGi9/n2JvKqic76uWSi6IewWz6b8SLbWKB82Oy1rBtGheb8w8jeZ69h/FLBnilF+ZGUzfosc945FF7nRZJLSOHNbsOY7+rKPBuRPsy8/LM0AM+/JcF7PDdj3JPPCRLbCcNs8hHWHcNsctGFNCz8HbmdfYgGU/h7Zk1GuqPlnENYdVLy3JkfCekwdfEGAPZdGs4/ZB6Cj89lMgeOs+X3uzfWKye0nadLd4z8XwNcXtKPLevTohYA4NaqQahhUpG9J4SEf0WC7yI0K5See5QNM9n4+xHmIbHzgkMFNjGh047P6yfUJlloMJn/OmgyYivMR0mOhDRi7WEY2DHXAEN7fhcC+8VeCHwqer6HPFiOWAkAuPSAOwLF656hCL8CuLqE14e5MNWmHcXRkA+oMoXfPQkAVp8NhekkH1FNRR+bRzhzfJh7rOW6GTnbcrPKrnDuIXI/GyYNwKdTjIi+t1v+ToW8/D+sC0ZhRJXz4uDggHXD2uHdNuYafmsF/wjQT8/RMttUahOXyIMiUTIeW7igb1/pUqAqA1VFyVAV8mbtazZ6DSJ/RCvZGvXx8sY5NKhXW662tsNccPt+0fkEE2RDnigZ2Tm5MGoj3VA+L8qIkvErJQct/rsjuaKMWNStg3uzJKds56Xt2kB8jklm1ymKUoofcge3EHz4FKnwfl55TIC5UQWp69+r2I29H/Tra4/99iYSWohGFRkB1YmGhgbijsrX0WA8nPuwWaeqKcL2yjaJ73a+pUz3aYqihLpeyIphn3kKx6yWN0qG2eQjxT5edqKvclJpA9woGfLyV2b6UwYeHh4lJh4voXRRrlw5ZGVlqdsMAkFqatSogW/fvhXZ8Xx9feHgIHyip6urKzZskD+Nb3HD2NgYCQny9cwTCATpUSjT39+MOLG8bt26IrSE8LdBxHLJo39/2XuKSwOc/2pRimVJcMRy4YnjyqRw6EpVQsQy4W9AVZmKlQXpYS4F5ObmQktL9ZMFihI9PT2kpqaq2wwCodhRUhLFqBpZvofS+J2VtM8UFham0jTNhJKNlpYWcnNzxdZ59+4dGjRooFI7iEsGDw9c/4faxvw9A58TMgXKWq+/gyeLmDBFVRZcwy+3nqwjPIEhNXAn9HtMR8p1D/YdAI5dfwTXPWfw/ZxgaJ+Bi3fj+MoJ0NEqC9M+sxF7aRv0e0wXeuE3NzdHVFSUyj+Huvj06ZPE1Ool7aZIkJ3oTb3Y5SoLruKXm/AQYzcj4tHN0gThUamwMtdD5flX5D5mrMcQLD37AmsHNoX/8x+wb8bE/jWb4YfVA5piYjf+6DhfEzLQZtUVxOwYjNvvYtClAZOrquEifySkZ7P1Ek/MZRMTAMDnmGTUNjMQa0utcTvx5QAzAdjIkesvqqGhgZhCs9hNxx9ArIhkBCmZudhz/RUm9WiMHZdfwuPKCwBAnJcTfiZloqoh9xof+OonbP8RnjvrRngUKAro1sgclSYW3eRCDlWqVMGvX7/E1onbK9qftfaM4/i8g3+yas7vfDSa54uDk7qiS8MqsHO7ioAFzP+skovys8AV5sKFC+jXj5kYyElwYjJiG+KPzWbfC3Pj5Vd0aFAV1cZ44PP+qag9fleRXgtTA3eK3Jablw+tQjkPrj56ja7NLVHPYTF+nN/EV0evYIJ7ymXZshMa9FmA5EvMfVS/t2z+2upGVr/kWy+/oJaZAWqZGYj0Q/758yeqVpU9i6gsEMHMg7hMN5KoPP8qAgMD0eStJxLSc9B61TVEbuqLyjOZeL7R25kwZc1XXEHYql4wm3FGKTYDQNpdb3yPTsA0t0PwXjkZNe2mIfXOIej9bwxS7/BPCqnYeTRSA3ci6MkbDFqyB14LnDDMpg0shi7Gh1P/Qb8H19Uk5boHezLLirgLiqyMWnsQ524/Q3rIYQQ+egW/oIfo2soKjj07oG6/Gfh0YQcqdh4Nl4HWePDyPe4fWoMKneRPFVpUtG/VHLfPHUZicgqMDJi4uNo1muLlzfOwrFsb2jWYWeU5315AqzoTJjHnyzOYNemC1zfPoXorZnJX9mcmQkt7+xF44M8kv9Cu1aLw4QhKJvvjA1y/8xD9xs3BgU3LMbz/v6jZ3g5fHwSgnAU38kPWh/vQqceE08p8cQlmHYdgz8qZqGpmgu7O85DxXDCZTvmmTCQNXsEsC1UWXAVN06AoCj9XtUdyVh4MyjHJC6queIDb05uhngl/5ATz5fcBMIJZWfCmtE48oVi0AF7BHHdgvEL7Gu95AxeeRCLOS/7Jg5UmHkEby2q4smwQktKzYVhBBwBg4rQDDzY6wcKcP869cUFc5KQzy1DLeROeeEyB5Tgm8kGiHxMdqduC/bjpxnw2w0FMNIF47ymoO+UAHm0YjoYzvQEAcYeY5BfWK/0QvJKJYW0ymklmIU4wA4DpRB/EejmLrcN+RgUFc8qVrcjMzsWdlx/xvyb1YD5wkdCJcj3m7cKj8E98GQFlxWTENtA0jZQrW2DQex6SL29m3zm8/x6LNhPd0Kd9YxxbNhr6vebCyNAAkUcWISktEy5bTuH0yjEwtHfF4z1zYVGtEvYF3McEO+Z8NrBjIm0o8/5WWDBXH7Ic308LRsUSB0cwx+4ejoDn39GiljG+xqej/7ZgxOxyBAAkZ+bCoDwz2rzsRUV4eXkh/tAkVBrrhbiDE/G/5adxZzX/uZ+Zk4eakw8g7uBEmIwRHrFJHhSZyMcRzHH7xqDezON4uHYgGs1lovJwYpfbrPVH0FImZrfJ+IMKJ4HiQHyYxXDw/jf8zuefQeoe/Am33scLrc/JjW5cQRuRm5jZttfmdmPFMgDsGCHbDHFpqV7ZGBe2zoOhni4rkguLZV5sWjdCynUPDLNhQnt9OMUk6Ei57oGU6x5YN3GASuwEgNqDuVkJOQK9/rAlWLH/gqgmLLZt/4HXkglw7MlcwD5dYNJyp931xpbZTrh/SLbMQupk+rgRAMCKZYARx5Z1a7PLwoh5eRuVjI2Q/TmUFcsA8MD/GHRqq+b/RRBOj/+1Q9aH+xjevyD71QNG/GZ9uI+sD/dx4cBWgTYx905joG0ntG3SAHeP84ew1G0mOhrPzYh4dN0SglNPuQkoEjNy0eq/WyLb2Fn9v737jq/p/AM4/jl2aSVClpotokrpMKpLiyAoIja1xa5Vq1RUW6QxihoxSlXtrVZo++s0SqkqSo0aWRKJUTPO749zz8m9yV3ZkXzfr9d95Z7nPPec58Zx8r3PfZ7nqy1tpwfLAMPqlbIIlp+c4Nya2K9P3m2xHXvrLvcTHnLt1j2O/JuyjIigpUMOO3KOrjO2Jtv32Vb76dht8RmSmL3yhZEpS6SyYK+2xJjv5MTe+enbjzFh3SGu377P6FXJ2xToq32Q1YNlgKvLB1sEyyW6zEr2uvNfvoe7SxFi140zgmWA74J74RbwUbL6/8ztSYmijxH9RT8jWAbYGxSAe/fky4Pa4jNstREsvzAm/Tpt7ClcqACNa1WhcKECyYLlj77Uftf9WthOVGOe3OTegwTj+e9nI22+Rg+SzYNlgEqlPYjbHsKK8d2MsmtxWj6EYk8UZm1Qd6O8YiltjHvvZnUp1jx5Jt2kXHwH8UzH8azYvd8ou37rNs92/oARc6wv+2rNxbUf4tp0JM+88zGuTROXOb1+6w6uTUcSc/0WL/a2nXylWY3SlHQtzMsVPIxgGcC1cAE8B6wEIDQ0cenD6CVacrOkwTJA4YL5jP0ZbdfhxPX8i7ebQsyN27wx0n7ykDOfdaTEE4UsEv0A7BnX3FiWce1a53/3aZGrA2bvkbuYsOUkZcaE4T1yl5HpZuquM9Sr5PwyQNXLWPYyvFbJw0bNrGPeq6wb2Drl61s78sHCTbj4DmJEx0a4+A4yzuviO4i/V33MxF4tHBwhZ9ttZW1lW0GzLeYBtMhc5r3KOt/X69h9zQtVLIc3WOtt1nVc/Bv/RN9iyJpjeI/ciffInbgVKcDluDs2X2PN8DdLW2xfnviyjZqJPAev45/IG3gOXmc83IoUpNTQDRQrUoAaZRxnDbWmYY3ybDuoZaMs3nG68XNOCgPmiesO4NFrMUOb1sCj12I8ei3mUuxNPHo5v8RmYP3KePT9iiMXYvDuv4L+S35mypajzNtzgqKP5WdK+5oOj7H3jwvJyq4uH5yi92IeQDvDPIC2xSNQW3s3/j9tHOjwr37l8OTWRnlWGfeO429QWplSbJfoNMMYxvBM/wV0nua4gyW1rm2dYnfbmvjds7lyNY7+IV/R9SMt0CvdaiThV+MJGdiG+JvOT9aO+yaY8BgtkC/efDR9QlZSpq2WzKR40SIcWpi6JGDmAXR2oa+b3DlY+wA3NHQnMatHU6nXZynO2mfOPIDODLk6YA4PbsTFKb6EBzcyHhXG7eHyVF+LANrciy++SMJDy2Es+pAMIFW9MCnVaKDWU1yhhZYApOjr3a3W6zVlGeUDRtMpaKExztictSA6rT7s3ZL43bMZ4P+m0ZOtP5Zsc7yoP2hDSpL+1J93Gf85u35NWYCZlfYf/oOHDx/yWDltsotvvVeMYRiO6D3JRSrUNLb3HTrKb38c51K47Z4Xkb66DQviyZpNaNt/NLdP/2I1aDaXkGTN06Q9ypcirH97BRAe3JiLUxoRHtzYeFQYF8Zn7aoZAXRSh6/codb0w1yITQyqzXuUy0zcZ7e9ushZAVye2ZrIWQHGQy9PLf3r96Q/a1b0ZnSA/d9jUhMCarF7XAv6NqxK1KKeRC3qydZRzYha1JNyA5bhNzl5L7Y1u8c0IWp+Z8LndmJuj1eImt+ZqPmdbdY/dCaCh6qKhykbav3nyibrUe4x2/oa3HpPsnvbjy223QI+wnds8m8HzXuSU9KrDBC14B2OXIghwrR+9bTOL7P/TBRRC97BI/DLDA2c6/TThhp4t0r+NXyj4bOZs+F7DpxI/kFDt3CQH2CZFfD5p7w4/rnt5Fj/XLnK76cvGglEXP1G4DdyrsV+c892n8y58MTVRsx7lJ3pXdbpf8+WjethbMfu1IbiuDxuPXmMLXHfBBP3TTAxW6cQOqKDsQ1QK9B6VkWAqqM3EvbnZa7fvm+U9Vj4E2ejbnAu+gYlB62i5ouJQ/bORsZz5Hw02w+fB8C9xwLceyyw2J+RYlaP5sjZCKJXacNdZvRpzL6Tl4xU147oPcnefbXhQ93nf4dn4FJjf61a6Zcgyx4Zw5wC+qS/hvXfYs+339mtmxG/V2fG6JifN6X1Uyo9x3idcq1NzZpaYJjR7c5s+fLlIyEhwW4d8/dz6dIlSpcubac2NKhfn7A9KUtPLFInpddj3759WbBggZ3asGHDBlq10oZEpXYMs/mkv8cKFeTOXfszzM3b+CiMYQ54uSLzer2R6mOdfy6QmjVrpnkMM0C+fHlJSLCf/CHl/4ffImzPXmPb0XVmfnxHY5hTIj0m/Tlqe548eYx7YFrGMLu2+5S8efM6dU6w/J1VqvC0wyQ25vUzcgxzaphP+nP03t1cXYi5ZpkIyJH0/puaHmOYDxw4QO3ate3Wbe3vz7r16Tf0SCb9pVBISAgjRqQ+paOw7v79++TPnz+rmyGEMPPKK6/w88/207BnR/YSlwiRUa5evUqJEqnP3CiyN5n0Z4ePj0+yMgmW06ZgwYJWyyVYFsK6H35I/xTVjuiJRh7FYFnkLI89lrKhDFnJWrCs93qLnC3XB8ynTp3K6ibkOHfv3s3qJgjxSHn99dcz/ZxlypSxuz+9lmnKCrm557l58+ZZ3YQUe9QzmzoacpdWZ8/aH0qSmYKCgtJ8jEc17srVAXOVKlXs7n+U/2BkJ1u2bEFRlBQ9jh8/ntXNFiJbS83/q6T3tPz58tmt5+xxbMnIe2i7du1s7luzZk2qjxsbG+uwzv379x3WSW/O/vtu27YtXf7903JdpfR9pWV/dpHWdjZt2tTmvqeeeoqpU6em+Jjp9W+rr/cOzgXMjo5XuXLlDL2mMkqOH8McEdzIqXpephUxLl68SIH1jpcI8hyyMU3tepTd/Mm5SSJ6YhF98oSeDXDy8u2M6eKHi+8gds8cRu0q5Y3X1A2czC8LtMkCqU2oIkR2ED61oVP1vEeFAVCngjub+jpe1sxrhLYiQ9TcjpQctIpiRQpyfEorak/YyqqB9Sjv/oRF/Z4Lf2Jx71ctyjz6fw1of9iGNHiaUY0qUWl8GH9P0tp87PJ1frtwje51yya2870dhCeZoKi35dr68U69Vz1RR+Ss1k7V15M/uT1RmFPTHPcaV/9gB1citBVknJ1cpi91FzW3o4OaGv135/ZEYU6FOF5BpPqEXUabUqvBc2W4d/8+64dYXlM+w9dwalpbPPouJ2p+F54btY4/pmptenrIKkJ7vUb9qomZ0fR67oFfsn37dpo2bUpUaDd8hq7k1IwOePddxgDfqozzfxGvwGVsfq8xzYN3EBXaDY8+Sx0u4+Xe23ZegJivHU8ILd7RLHHNQueWG3PvbX8dX2fF70i+nro1Lk0Sr6vrexxPDFRVFZeGlitSpfRcYzr5Mrqj4/vJ9uul6NBBW1Zudl8/Or1ZjUq9Z/P3wsTzV+s/j2Nz+9F+6joWv/s2ZbrNJGaVtoTdk+9M5/KX2jn3HDlL/epPcfD0ZWpVehK3ds4H7PWqP83GIMvEORW6BnNm2Ujc/IOI3RBElV7T+GuRdk3oZeb0smKtJjh93vQgY5gB71G78R61O1mZTg+s9WDZa+gmvIZusqjvNXQTf17O2OVXsjtbwbL5EnA3/7O+ZqyeOntMFz9j2zxYnrk6zAiWhcgJSo7eYzx05cZ9a1FHHz6wqW9NvN/bYTyccWV2e45P0Vbb2D+xuUWwfOvuA2bsPJ4sWDYX/mkTRjWqBGAEywDVnixK97plib11j/sJDzl79RZ7hlpPPNGybuI3dRW6JwY8+lJqf56PxKfXjGSvm7jpGF7vbrBYlhNgy++XjOd6L5N5sOwRmHgP8uy7DI/AZUbZ0Q+1gL5lneRzUwA8u8wEoO9c67/foA2/4zlgpZH8Qbf58L/J22QWLHv0TUyh7dlvOR59tcefF2M5OtG5Tht7ihTIkyxYhsTfS5RpKTk9WPbou5x/Zra3CJbN6wHcvHnTCIBPzdCCrPD5XRnnry1nGbGgK7UreBIV2o0Jaw9yae47ePRZyqAvnFse1BklOk03fpboNN0IqvVg+bUgy44pjz5fGD/15+nN1W84rn7DbZbpwa4eLNfp/YlFXT041n/a6inVj5n0XOb0b0rMg2XXZonrMxdrPgrXZiONMr+iif93ihTS5gyZB8sAx+Zqa3qvGhVAkUIFjGC5ePtgI1gGaFDjKRQFalVKeSrqpMEywJll2nn0wFgPlkPW/pAsWDavl53kc1wl5wif6suF2NuUdXvM2LbF2/Uxfp/QiNIjtnAxRMvoFzGjZaa081EStv8YDWtX4+lSnoCWje/no6d4pbr1P1j2DGnnXI+cEI+KK1MaABhrt/8ddYvzH9lOGKT34PqM13qdrfXqOqtIwXwMbfxsql6rcyuipdl9qkQRp+o/X6Gk9rqAj3itajkAqpbzJDruVrK6E1pWY0LLaiQ8VHn1kzCW9KhNJa+ivP18KavH9p++i32nE3tqI+NvEznf+eXVgtf/ysuVtWPP798E/0/Ws2GsZU93kP/zBPk/r7Xpw29Y0vtVKnm70OIF6+O9/WeEJW/TvNQvY5dezAPj9DCxjfbNR3olith79DyHzoTzlJeWDOfqimFG8GxuTndtbL9Hny+ICu3OFdO/d1Rod349HZEubTH3wLR0YNJshdZSfet8yngB8NagEL6dPYL4MK1jKD5sNrdu36XIY9YnwevHTHj4kFqBU/ny/W5ULuNp8zwv9gnmnytX+bhX4rru17amfJiGLXrgnNlGtMn8+RuplWt6mPXgWA+WvUftZsDKY3iP2s13p5InEvh9gtYroAfLwrqGtasBcGRl4n/c2lUr2KwfullbDUBPmnL7buaPBRQis+XNo/UyVfIowtZjkcl6nc0dvRTPKVNvb/inTSg1Slv/3VGvs94reuRfyzG4Y9Y4N6zt78ibxjnqT/+JmFuW6zq/9PH3Do+x9n2tpzJ23Tg2B2nJQNzbfmw3s13ePAo/jW1IJa+ieA/ZYPQ6Jx0tuGFYI7aPbkrUAi1o8nR5DO9+WjIOj8Bl/Hgy3G7bRrZ+mU3vaz2wnl1mJguWk7Xpg6ZU8nbBa+BKq73OABuGNmT7qCZGgOrp8hje/bXU3f2W/GTR85wR7txPwGe41gvp0Xc59x4krhU9e1fq54F0mKV9YPv2uJaifdn/Eidp6Ukk0qJ+9XKMbP0yB6Z1N3qXraleVkv7HhWqJefKlycxZHm5olea25FUvrx5GN3JsiPN1W84Jf3H2OwNXjZe6w3/dnby1bVsBcvm8ubJw4EFo6hcxpNiTUfYPM+hUC2gHdAyMcDUe5bNe5lTYuq6n2n8QWKq+b1Hz1G8ve2U3Gk1b5uWRMnNP8iiPOl2dpVrAuakwqf68nmHaoRP9eVNnxIUH6qlqw2Y/6uDV+Zu+rhke/LZWGJn5uowmr3yHJt/PML59VNxa/wujxXMb6TQzojMg0JkN82reXJlSgOj9znpBLXqpVwsti9N1ZItOeppfqOyFyt++Ychy/cbZXH/3WPpD2ecalclz8d5q7I7AFNbV6Vq0F6L/b+9X8/q6zb98pfd40aved+p8wOEz/Qn4jPtoSjJkynUMAVQRv152le/UQu68lplb+6begg37bM/Cz9y+RCn2xQxpwORn3cwUg47bNNc7YPCvB6vEjW/i9GmjHDj9n2LMcwF8iX+SZ+08TCVhq1O1XEfPFT56qe/mbhO+7DV9Q0f4v6znxgnta6uGGY8PExDZqaftd3TmtFGd7IcQhO3fRpXNkwmbvs04rZPY8oP0QAUa/Ruup/72jchxnkg+WovcduCrW7HbQsmblsw87c4Hi4THa9921O8fTCjAl5h54fa9Tpo/g72/J6xq3H0a1aH1f87SuyGIHx6aJkM9bHKj0LQnOMn/Ymsl9qMSTLpTwjbnJ2klpTvgmMcOXoMgIgQv1SfX5/wJzJH9ILk40JTY9JxF2bN0oYNRC/sni7HBNjn5f9ILmmXk8WuHpUux/m58Ksp+re9tnFiupy3TJdgbtxMPpwrI0mmPyGEyCHWrl1Lmzbpl9paCCGym9u3b2dJQhtZJUMIIR5xelZSCZaFEDmdHiz//fffWdySRBIwi2zr6NGjWd0EIbKUecpdZ7NjpWeaXk/PrBtLKrKvwMDArG6CsGH+/PlZ3QSnjBtneyKwuUqVtGUv8+TJ+nBVhmQIp13fPctq+eYfjzB8zjrOrP7IovxiVCxVu0ykRsXSHDl9kfhdnxn7ivpq611f3/O5UbZi9z7y581L2/rWkzdM/nI7Y96xP+by20MnaDX6c/Llzcv9Bw+cel9CZCSXJ57g4YO7nJ3Z3qJcn6hlS6mBX3Mt/jqFCxe2KP9rUhP+Cr/OaxVLON0Gr2FbWN6rNg2reOI5dDNgO6mT96jdlHi8AMfG10u2r92iQxy5FE9ox+q8USlxspue+EnPCiYyTr58+XiQwfc2Z84RERGBl1f6r1QhNOW8i3Pv3n3+nO/8XB7/SSv537HzxKwZg1ubxPWhFUVxKnGMqoKiwENVpe57S3m1ShmWhP2eqvZ38a3FlatxrP+wp9OvebqsJMe2AAAYRElEQVTjh8Te+I9rW6fg2sxy/PXx48epWrUqV1cMtXuMTtM28/2xf8mbR+HfJQPpNfsbNv7qfCpuGcMs0oWtgDk1rAXM6a1ogwEZdmwhnJXWyVrugV9abEfOaJGm4zkKmFNj8G8urFmzhshZAQQu3c+CbrXxenc9EZ9ZLt3m9e56nixWmENBTfAcvC7dzp/TtKpZntBer9n8UDV+7W88XsOPkJAQouZ2wHPAKiI/b0//pb8yt9vLRr2DZ69S86kSdA/9iS/6vIpHf8ul8aJDu1kkggGMZfs8ApcZz8/V6E+tWrWMSYJJl5aztj6zvax/wrHYtWPT9Ho9YHY2WLZ5nA4hqXpd3La0rRGdNGB2NnOnLXpGT0dkDLPIUHX6TDaeu5iW2pmweCt1+6Z9UfX24xfg0nAgbo20ANvFtHLG1p+P8lyXzE2ZKUR66rHgf+m2Tu/5mFt4DduC17AtRtn2Y/bXJXaGng314x2nnX7Ngm61AZIFy3rZoaDUJWLJTUJ7vQbYTkAyqc1LTJuWmEwj8nPt2wvzYBmg5lPatxBf9LGd7TFqQVeLh3m5rnbt2pavCe1m8RCZo3jbycYDYOgC5zKCJnUuMo7iHadZpCL/5uBpI3FLRpny9R6KNR9Nseaj0/W4JTrN4MNVP1lsZ4RclelPZIx9oYnprPVhFxN7Nmdiz+a4NHrXYihGSq2aZDlWLt60RF3zV6rT/JXqqT6uEFltSeAb6XascsWLEDHdMsmSXzXvNB9XT/j0fpOKTr/m+OV4Os7/icjrd3i+rBv93qzIoh/OsOXdemluT25SZtDX9Kjnw5tVSvLGM95UHrGGkyFtbdZ/buxm/F8qy75/otn5XkOaT9vD1uENHJ6nWfAOto1M/BDTdd53LOv3JgBffH+K7vUss7YevxRL+8/CiIy/zQvlS9CvYVUWfvsX20amfolC4ZyYNWMstmcEpu7DZ3lPV2K+Hs7U9b8YZU1rVmTv0fPUr14uLU20a3THBozu6PiaTKmkwzQcDdtILYcBs6IoS4BmQJSqqlVNZUFAbyDaVG2sqqrbTfvGAD2BBGCwqqq7TOWNgc+AvMAiVVWnpO9bEdlRWoJlIcSjx8e7KF/2rkv1MsWMMm/XzF8e6lHWdf733LmfQFDrF/Hst5zIeV2MYPnDDYeZs9syi1+1MZtY3u91wq/9R5B/DQA+aKX99BywilOf+uNauIDVcx34J4qvfjrNeyt+JXzeOyzr9ybno2/w29loLsbe5N+rN1EUxajvU9KV5QPrU6Ns4hj6ksW0cfYefZZKj/MjZFTruhbbGRks5wTODMlYCjS2Uj5DVdUapoceLFcB2gPPml4zV1GUvIqi5AU+B5oAVYAOprriEVfafzTl24yluN9Q9vx2wij3aKaNmRoXuilNx3dJkrzEfDvpPiEeBR59lxsPczfv3KfqyJSP6935ZwQA765MnJxjPjzDfJiGIz4TvuXNGb9QekyYMRwDoOz7e4iIv2vzdbNna4kwPAevI18exSJYBqhZ3jIbnoxftm9Z33rGcIzIeZbDMj7wf4Go+V2MyZUe/VdybHJLapRxo0n1UkY9fThG5OftcS1cINn4ZdDSd0ct6ErnVysaWRMByrk/QUDtp/jA/0XKlHichw+1r+rde39Bvjx5LIJlgJpPewCJY5ll/HLGmLr2R4uU8cXbTubPC1HGEA1r5syxTBy245CW+XPg/J2Jx+k4jbI9ZicbppERzOfN6cMzUjNMo0SnGcYjafmde5YTVu/dT0h9g804NelPUZRywLYkPcw3VVUNSVJvDICqqpNN27uAINPuIFVVG1mrZ4tM+ns0Xb9+naJFi6brMcuWLcuFCxfS9ZhC5CR+fn5s3y7Z94QQqZM3b14SEtInuLTlwIED1KpVK0PPkRYZNelvoKIofyiKskRRFL074UngolmdS6YyW+XWGttHUZTfFEX5LTo62loVkQ2Ehoba3JfewTIgwbIQDlgLluX/jRAiqW7dulktT0hIoEKFChl6bmvB8uuvv54h50rPNekh9QHzPOBpoAYQDqRbH76qqqGqqr6kqupL7u7u6XVYkc769OmT1U0QQjhQtmzZDDv2jz/+mGHHFkJknKVLl9rcd+bMGYYPT/0ydPaULFnSavkPP/yQIedL797yVAXMqqpGqqqaoKrqQ2AhoH9kuAyUNqtaylRmq1w8QswnfqRn3cw8lhAZKSuuVT2FbEaw935ee01b+uzcuXMZdn4hcqsNGzbY3Z+R95pp06ZRoID1SaJpceXKFbv7bQXU2UVqxzB7q6oabno+FKitqmp7RVGeBb5GC6BLAnuBioAC/A3URwuUDwIdVVU9jh0yhjljXA+bzZ6DJ+g0cRGR26ZRPmAMX03oxSvVnk5W18V3MPGmhCVxN//D9XHLrGNFGw4CwLdhA1b3f83huYu1sr928vWwOXb3gzZpwMVXO++/m6bh+kRhB6+AovX7OawjRNTn7fEcuJpjn7Sg2tjNfNr+Jd5b9RuRc9oBsO7gBQJqar22ngNXA/Dsk658O0ZLAlJ+2HrOTdfWH/YYsAqA6EVapqtf/47g5UpetJ6+k3OR17lz/wGHp7Zj08GzfLLxEBFx/xG1sAcAr03YyI8TWwHg3msx4HyiET3r3q+//spTB4Id1vd8d6PxPOJTa/O7LdX45Ecirt0CIHzym061yXvMd07VEyI3mDJlCoEVbjmsVyxgkvE8dvUoOzU1bu203Af58+XlyoLuDuu791ykHXvVe3h0msbxuf2YtOpHZgU24r+79+kUspGr8f/h4VqE9WPb8O0f53nruXIAFO8QQszKEew9eo761cvj1v5TQBsGEbNhosNzu7ZITI0d/43j+1SbmbvZHbYHgJgVziUxKd5putGmq2ved1i/WOsP0zaGWVGUlcCvgI+iKJcURekJBCuKckxRlD+AN4GhAKYAeA3wF7ATGGDqiX4ADAR2ASeANY6CZZGxGtR8hsht2kiac+smWw2WS7Z4j/jds3D3G8qegyeSBcvm9GC57rtzAXAP+BAAN/8gi3rXNtr+j9S2beIao+UDtBmzLr6DjOB49wHtkjH/ZJ00WHZp0N943jkocZy19FILZ0XOaYdH0UJEzmnHO68+bQTLgBEsf38igicK5afFC6WNYBkwgmVrXq6kpRFeP6wxh6e25a/pHSmUPy/t61bkj0/b88Rj+fHovQTACJaTMl+5wvz5+t+TJynRg+U3puwF4MlhWoY/ryHayjU9lxzQ3u9n2rmsBcvepuB778loWs0/wNuf7+fIWMsPxiXHfm/1+fojkVbfgxC5nR4s1x22AAD3dh8D4NbmIwC6hqzjQcJDrq0bDyQGy5UDtcy4xdtr/7fX/HjcYlunB8vPDvsaAA/Th+51+85YbEcv7mW8JmrFcNxdCjMrULufFS6Yn43vt+XH4G6sH9sGwAiWV3x/jJiVIwCoX728xbn1YNmnmxa8F2upvYc1/ztqUS9us/Ze478JZs+hU9TpP53GI+cBUKmz9kFBz/anqiprh/iS1PnIOABKdJ7BniOJ33SV6Gy5coYeLPv00gJoN9MHkTU/HLOod239B8nOYc5hwKyqagdVVb1VVc2vqmopVVUXq6raRVXVaqqqPqeq6tt6b7Op/seqqj6tqqqPqqo7zMq3q6paybTvY0fnFVnvymbtE2P09hk0qPmMw/pu/kGcvBiFm38QCQ8f4uYfxNRefsmCZmfUez5xsfz43dqSVQOmrbBaVw+SXRr058rWGcb21p+OpPi8Qjij3jNenAnxJ7RHXceVnfTPrC5GD7MjT44OI3yqLyPW/wXAwFXHrNbzGrKJUxE3mLT1OAkPVSNY9hqyicU9rM9Uj7pxF++Ru/ju1FXCgxvhPXIX9Su7s7FvLQ5eiLPZplLv/48rn9TjvY2nABi05oTNukLkdm5tPuLkxWiCvtpLwkPVCJbd2nzEshEB5MtrGZ4Vbx9MdPwtIzgu3j6Y6//dpXj7YGJWjUx2fI9ei7l2864RHHv0WkxAnQp49FpMlOlbr9TqVK+a3f3FWo4nKu6mESwXazme+Jt3jO2kGrzow8l/I9n313m6TVnBqeXjLFJjW+vwWhR2hH8i4/jzQjRXvxpKgxpa4F6i8wyufpU8cYlbwCSi424ZwbJbwCTib90xtp0hmf6EYfLyHYzp0oSPln7DuG5NmbEqjElLvyF250yLoRm2xG4Islre2y/lS8h88b72CVkPlgFOr/7Eat34PXPt/hQire7eT6DM0HVEzmmH58DVlHiiEFdv3LHofXbWicvXCL92iwlrD/LjxFZ49F5C1MIexk979Mx7l6c0BCCkdRWL8qQiZrY0no9v/qxT7fN4oiDhZsM/bD3XXfmkHgCXPtYyF37ayseiXAiRXOzaxCEJQZ3rO6xvLSgG6NXoBavltoLilATL83ccom+TF43hFwCV+87l5Pz+dl93bZP1ILR309pWywHitk21u51Ur4Y1rJZbC5YBYtdZD9Z7N6lp9zzm0rKsnMhBfvrjDKv3HsTFdzBzN34PwND2DUkwLVjvKFh2VtzN2zb3rVmzJl3OYY0zY/WFsMZz4GpOhMcbwXHknHaMblbVCJ5TwqP3Ehp+tJlbdx9wKeYmAFELezD8y5+d7l3OKCXNhngIIbIPn8CM6fypMnyV3f19m7wIYATLANHx/6XLuXvP0JIX2YsJMlufz+wnWnNq0l9WkUl/GeN62GzHlZykT/oTIieI+rx9uh0r6aS/1NIn/QkhREaJXfVeml6vT/p71Nmb9CcBsxBC2FGwYEHu3rWdFloIIUTOkFGZ/kQO07+//XFJQuQ2jz/+uATLQgghJGAWiebOlUlyQpi7efNmVjdBCCEyjb0sgLmdrJIhAChQoAD37t3L6mYIkS0oipLhE0Uz4xxCiNzt6ZIlODitm9P13y4IsWaT/My5dQix2M5t9zAZw5yLxO+cYbXctckwBvi/wZvP+xAwPpS4HdNxaTyUCRMmMPxV92T131+wkY8DW+HiO4inSrrz+9IPKNpwYEY3X4h0EzWnrdXyZ8ZsYd8HTag0chORs9uwev95Bi0/wBOFC3F2RvLX3Lr7gAV7TxD67QlOhrTFPfBLq8eNXmg769atu/dZ8t1JBjWuhnvvLwCI3x5isz5A6/GLWD+pFy5+1v+wCSEEaMFvnRFfcPpKLK1e9mHRoGbGvuIdteRl0SuG4W7Kihfz9XAApq7/hVGt6+LReTpRX2mZ9fSA+do621nzbt25R5FCWlrtYgGPXsoNe2OYpYdZELdjutXn1oJlgI8Dtcxg5mskC5ETnJj8NgCRs7XMVu1qlwOwGiwDFCmYj2F+1RjmZ38hf3uKFMzPoMYpe/36Sb0cVxJCCGBfiOUH9uIdp9HuNW0Ndz1Ajvl6OKevxFK84zT8XqrA8mEtAIxg2Vl6sJwTyRjmXMZ32Cx2HzyBa5Nh/HfnHm8OnoFXC/s56gPe19JV7jn4F5OXb2fJtp+MffM3/S9D2ytERvk74jp1J+0EYP8/V4m5eZfKozfbrH8qPJ6XJ2j795+JIubGHXyGO792eJkBywF4LUhb6zPs2CW79b1ajTWljNV6dbp8nNh7fejvi06fVwghzMV8PZy5/ZoYwbKuYkk3Yr4ezvCWdRweo2SnYPYeOUvdYQsB6Bqy3th3+MyV9G1wNiEBcy5z6NS/tP1gIXE7plO4UAF+P32RiM1aRp2lO34l/lbyRcS/O3wSgC0/H2XK8h30aPaqsa9vyzcyp+FCpLNKXkVp9WJpVBVqP12CNz7ZzckpLXj9k10AeA5aa1Hfx9uFC9E3AFiz7yzPvLeWU9Os9zxbc+d+Ah59lnLqShzefZfRsFopPPosBWD8mgP0Dv3eov79BwkEfLCIfXO1YRdbf0lMf+03UiboCiHSbv6OQ8nKajzlafc1NQfP5869B9y6c4+TF6Nxa/MJy0a0psXEFXh1mMoLFUpmVHOzlIxhzkUGDx7MJL/yDuu5NE5MLXk9bI7D+v9GxlK18wdpapsQmSlPnjxEzApwWM9jYGIPcvSCdxzWT80YZovXm8Yw58mTh2vbgu3Wrdh5ElGx8U4dVwiRO1mbwFe84zRivh5u8XPVSH8a1tDig6YTV/HNhORJnJwZw2wup41hloBZCCGEECKHatvW+jdhGzZswN/f36ljrFnj/PCzR5lM+hNCCCGEyIVyS7Cb0WQMsxBCCCGEEHZIwCyEEEIIkUsoimJ3f758MvjAGvmtCCGEEELkMDHLB1Nn1Ffsm9rZKPvyuz+5+uUgYpYPNsqKd5mFoijUrFiSnR92pFD+PMSuek/b1yGEmJUj+ObgaZrWrIhb+08z/X1kFzLpTwghhBAihzEPiu0p3mWWESA7ktMDZnuT/mRIhhBCCCFEDrb3jwsEb9zP8X+vGmVvjFtpUWfkF3sZviiMExev8mTXmfh/spavvjuW9FC5lgTMQgghhBA5UMkeWpKjjtO3MrJVbZ4tU8LYZx48qyoMal6TpXuP8kzpEty594ANY9vQ+c1qgDY0I7eTIRlCCCGEEDnMxA6vMtDvBbt1ineZZTx3ZlhGbh6Ska0DZkVRbgCnsrodItsrAVx1WEvkdnKdCGfIdSKcIddJzlRWVVV3azuy+yoZp2xF+kLoFEX5Ta4T4YhcJ8IZcp0IZ8h1kvvIGGYhhBBCCCHskIBZCCGEEEIIO7J7wBya1Q0QjwS5ToQz5DoRzpDrRDhDrpNcJltP+hNCCCGEECKrZfceZiGEEEIIIbJUtg2YFUVprCjKKUVRziiKMjqr2yMyl6IopRVF+U5RlL8URTmuKMq7pnI3RVHCFEU5bfpZzFSuKIoyy3S9/KEoygtmx+pqqn9aUZSuWfWeRMZQFCWvoii/K4qyzbRdXlGU/aZrYbWiKAVM5QVN22dM+8uZHWOMqfyUoiiNsuadiIykKIqroijrFEU5qSjKCUVRXpb7iTCnKMpQ09+bPxVFWakoSiG5nwhdtgyYFUXJC3wONAGqAB0URamSta0SmewBMFxV1SpAHWCA6RoYDexVVbUisNe0Ddq1UtH06APMAy3ABiYAtYFawAT9j6LIMd4FTphtTwVmqKpaAbgG9DSV9wSumcpnmOphuq7aA88CjYG5pnuQyFk+A3aqqloZqI52zcj9RACgKMqTwGDgJVVVqwJ50e4Lcj8RQDYNmNFuRGdUVT2rquo9YBXQIovbJDKRqqrhqqoeNj2/gfbH7Um062CZqdoyoKXpeQvgS1WzD3BVFMUbaASEqaoaq6rqNSAM7SYmcgBFUUoBTYFFpm0FeAtYZ6qS9BrRr511QH1T/RbAKlVV76qqeg44g3YPEjmEoiguwOvAYgBVVe+pqhqH3E+EpXzAY4qi5AMKA+HI/USYZNeA+Ungotn2JVOZyIVMX3U9D+wHPFVVDTftigA8Tc9tXTNyLeVsM4GRwEPTdnEgTlXVB6Zt839v41ow7Y831ZdrJOcrD0QDX5iG7yxSFKUIcj8RJqqqXgZCgH/RAuV44BByPxEm2TVgFgIARVEeB9YDQ1RVvW6+T9WWeJFlXnIpRVGaAVGqqh7K6raIbC8f8AIwT1XV54FbJA6/AOR+ktuZhta0QPtwVRIognx7IMxk14D5MlDabLuUqUzkIoqi5EcLlleoqrrBVBxp+moU088oU7mta0aupZzrFeBtRVHOow3begttnKqr6StVsPz3Nq4F034XIAa5RnKDS8AlVVX3m7bXoQXQcj8RugbAOVVVo1VVvQ9sQLvHyP1EANk3YD4IVDTNTi2ANoB+Sxa3SWQi01iwxcAJVVWnm+3aAugz07sCm83K3zHNbq8DxJu+at0F+CqKUszUg+BrKhOPOFVVx6iqWkpV1XJo94hvVVXtBHwHBJiqJb1G9GsnwFRfNZW3N816L4820etAJr0NkQlUVY0ALiqK4mMqqg/8hdxPRKJ/gTqKohQ2/f3RrxG5nwhA+5oq21FV9YGiKAPRbkR5gSWqqh7P4maJzPUK0AU4pijKEVPZWGAKsEZRlJ7ABaCtad92wA9tgsV/QHcAVVVjFUWZhPYhDOBDVVVjM+ctiCwyClilKMpHwO+YJnqZfi5XFOUMEIsWZKOq6nFFUdag/XF8AAxQVTUh85stMtggYIWpE+Ys2j0iD3I/EYCqqvsVRVkHHEa7D/yOls3vG+R+IpBMf0IIIYQQQtiVXYdkCCGEEEIIkS1IwCyEEEIIIYQdEjALIYQQQghhhwTMQgghhBBC2CEBsxBCCCGEEHZIwCyEEEIIIYQdEjALIYQQQghhhwTMQgghhBBC2PF/N4h2+3zguw4AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 864x864 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "\n",
    "\n",
    "dot_data = io.StringIO()\n",
    "\n",
    "plot_colomns = list(df_p.columns)\n",
    "plot_colomns.remove('OEE')\n",
    "plot_colomns.remove('Product')\n",
    "print(plot_colomns)\n",
    "export_graphviz(regressor, out_file=dot_data, feature_names=plot_colomns, class_names=[\"OEE\"], rounded=True, filled=True)\n",
    "\n",
    "filename = \"Decision_Tree.png\"\n",
    "pydotplus.graph_from_dot_data(dot_data.getvalue()).write_png(filename)\n",
    "\n",
    "plt.figure(figsize=(12,12))\n",
    "img = mpimg.imread(filename)\n",
    "imgplot = plt.imshow(img)\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Scaling the data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Support Vector Machine"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.804315810116353\n"
     ]
    }
   ],
   "source": [
    "\n",
    "svm_classifier = SVR(kernel='linear')         \n",
    "svm_classifier.fit(X_train_s, y_train)\n",
    "y_pred = svm_classifier.predict(X_test_s)\n",
    "\n",
    "print(\"Accuracy:\", r2_score(y_test, y_pred))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Random Forest"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Hyperparameter Tuning using Random Forest\n",
    "\n",
    "def random_forest_hpt():\n",
    "    n_estimators = [int(x) for x in np.linspace(start=100, stop=1000, num=10)]\n",
    "    # Number of features to consider at every split\n",
    "    criterion = [\"mse\", \"mae\"]\n",
    "    max_features = ['auto', 'log2', 'sqrt']\n",
    "    max_features.append(None)\n",
    "    # Maximum number of levels in tree\n",
    "    max_depth = [int(x) for x in np.linspace(10, 110, num=11)]\n",
    "    max_depth.append(None)\n",
    "    # Minimum number of samples required to split a node\n",
    "    min_samples_split = [2, 5, 10]\n",
    "    # Minimum number of samples required at each leaf node\n",
    "    min_samples_leaf = [1, 2, 4]\n",
    "    # Method of selecting samples for training each tree\n",
    "    bootstrap = [True, False]  # Create the random grid\n",
    "    random_grid = {'n_estimators': n_estimators,\n",
    "                    'criterion': criterion,\n",
    "                    'max_features': max_features,\n",
    "                   'max_depth': max_depth,\n",
    "                   'min_samples_split': min_samples_split,\n",
    "                   'min_samples_leaf': min_samples_leaf}\n",
    "    pprint(random_grid)\n",
    "\n",
    "    # Use the random grid to search for best hyperparameters\n",
    "    # First create the base model to tune\n",
    "    rf = RandomForestRegressor()\n",
    "    print(rf.get_params().keys())\n",
    "    # Random search of parameters, using 3 fold cross validation,\n",
    "    # search across 100 different combinations, and use all available cores\n",
    "    rf_random = RandomizedSearchCV(estimator=rf, param_distributions=random_grid, n_iter=100, cv=3, verbose=2,\n",
    "                                   random_state=42, n_jobs=-1)  # Fit the random search model\n",
    "    rf_random.fit(X_train, y_train)\n",
    "    print(rf_random.best_params_)\n",
    "\n",
    "# random_forest_hpt() "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.8866850503660298\n"
     ]
    }
   ],
   "source": [
    "# random_forest_regressor = RandomForestRegressor(n_estimators=100, random_state=0)\n",
    "random_forest_regressor = RandomForestRegressor(n_estimators=800, criterion='mae', max_features='auto', max_depth=80, \n",
    "                                                min_samples_split=5, min_samples_leaf=1, random_state=0)\n",
    "\n",
    "random_forest_regressor.fit(X_train_s, y_train)\n",
    "y_pred_t = random_forest_regressor.predict(X_test_s)\n",
    "print(\"Accuracy:\", r2_score(y_test, y_pred_t))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Neural network"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From /home/dulanj/Environments/py376/lib/python3.7/site-packages/tensorflow/python/ops/resource_variable_ops.py:435: colocate_with (from tensorflow.python.framework.ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Colocations handled automatically by placer.\n",
      "Input Dimention: 17\n",
      "Model: \"sequential_1\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "DenseLayer-1 (Dense)         (None, 192)               3456      \n",
      "_________________________________________________________________\n",
      "DenseLayer-2 (Dense)         (None, 64)                12352     \n",
      "_________________________________________________________________\n",
      "DenseLayer-3 (Dense)         (None, 16)                1040      \n",
      "_________________________________________________________________\n",
      "DenseLayer-4 (Dense)         (None, 1)                 17        \n",
      "=================================================================\n",
      "Total params: 16,865\n",
      "Trainable params: 16,865\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n",
      "None\n",
      "WARNING:tensorflow:From /home/dulanj/Environments/py376/lib/python3.7/site-packages/tensorflow/python/ops/math_ops.py:3066: to_int32 (from tensorflow.python.ops.math_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Use tf.cast instead.\n",
      "Accuracy: 0.9237331254578596\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model = Sequential()\n",
    "input_dim = len(X_train_s.columns)\n",
    "model.add(Dense(192, input_dim=input_dim, activation='relu', name=\"DenseLayer-1\"))\n",
    "model.add(Dense(64, activation='relu', name=\"DenseLayer-2\"))\n",
    "model.add(Dense(16, activation='relu', name=\"DenseLayer-3\"))\n",
    "model.add(Dense(1, activation='linear', name=\"DenseLayer-4\"))\n",
    "print(\"Input Dimention: {}\".format(input_dim))\n",
    "print(model.summary())\n",
    "\n",
    "# compile the keras model\n",
    "model.compile(loss='mean_absolute_error', optimizer='adam', metrics=['mean_absolute_error'])\n",
    "\n",
    "model.fit(X_train_s, y_train, epochs=100, batch_size=10, verbose=0, validation_split=0.2)\n",
    "\n",
    "y_pred = model.predict(X_test_s)\n",
    "print(\"Accuracy:\", r2_score(y_test, y_pred))\n",
    "\n",
    "# SAve model as image\n",
    "from keras.utils import plot_model\n",
    "plot_model(model, to_file='model.png',show_shapes=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Result: 0.336980920321386\n",
      "Actual: 0.590192043896313\n",
      "Error:-0.253211123574927\n",
      "Squred Error:0.06411587310207693\n"
     ]
    }
   ],
   "source": [
    "number = 3\n",
    "test_results(regressor, X_test_s.iloc[number], y_test.iloc[number])"
   ]
  },
  {
   "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": "3.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}