06-1. Adam, relu.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-11-27T16:41:01.694377Z",
     "start_time": "2019-11-27T16:40:54.210691Z"
    }
   },
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import seaborn as sns\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.preprocessing import LabelEncoder\n",
    "from sklearn.decomposition import PCA"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-11-27T16:41:26.091872Z",
     "start_time": "2019-11-27T16:41:07.516688Z"
    }
   },
   "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>beer_id</th>\n",
       "      <th>username</th>\n",
       "      <th>date</th>\n",
       "      <th>text</th>\n",
       "      <th>look</th>\n",
       "      <th>smell</th>\n",
       "      <th>taste</th>\n",
       "      <th>feel</th>\n",
       "      <th>overall</th>\n",
       "      <th>score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>271781</td>\n",
       "      <td>bluejacket74</td>\n",
       "      <td>2017-03-17</td>\n",
       "      <td>750 ml bottle, 2016 vintage, bottle #304 of...</td>\n",
       "      <td>4.00</td>\n",
       "      <td>4.00</td>\n",
       "      <td>4.00</td>\n",
       "      <td>4.25</td>\n",
       "      <td>4.00</td>\n",
       "      <td>4.03</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>125646</td>\n",
       "      <td>_dirty_</td>\n",
       "      <td>2017-12-21</td>\n",
       "      <td></td>\n",
       "      <td>4.50</td>\n",
       "      <td>4.50</td>\n",
       "      <td>4.50</td>\n",
       "      <td>4.50</td>\n",
       "      <td>4.50</td>\n",
       "      <td>4.50</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>125646</td>\n",
       "      <td>CJDUBYA</td>\n",
       "      <td>2017-12-21</td>\n",
       "      <td></td>\n",
       "      <td>4.75</td>\n",
       "      <td>4.75</td>\n",
       "      <td>4.75</td>\n",
       "      <td>4.75</td>\n",
       "      <td>4.75</td>\n",
       "      <td>4.75</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>125646</td>\n",
       "      <td>GratefulBeerGuy</td>\n",
       "      <td>2017-12-20</td>\n",
       "      <td>0% 16 oz can. Funny story: As I finally wal...</td>\n",
       "      <td>4.75</td>\n",
       "      <td>4.75</td>\n",
       "      <td>4.50</td>\n",
       "      <td>4.50</td>\n",
       "      <td>4.50</td>\n",
       "      <td>4.58</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>125646</td>\n",
       "      <td>LukeGude</td>\n",
       "      <td>2017-12-20</td>\n",
       "      <td>Classic TH NEIPA. Overflowing head and bouq...</td>\n",
       "      <td>4.25</td>\n",
       "      <td>4.50</td>\n",
       "      <td>4.25</td>\n",
       "      <td>4.25</td>\n",
       "      <td>4.25</td>\n",
       "      <td>4.31</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   beer_id         username        date  \\\n",
       "0   271781     bluejacket74  2017-03-17   \n",
       "1   125646          _dirty_  2017-12-21   \n",
       "2   125646          CJDUBYA  2017-12-21   \n",
       "3   125646  GratefulBeerGuy  2017-12-20   \n",
       "4   125646         LukeGude  2017-12-20   \n",
       "\n",
       "                                                text  look  smell  taste  \\\n",
       "0     750 ml bottle, 2016 vintage, bottle #304 of...  4.00   4.00   4.00   \n",
       "1                                                     4.50   4.50   4.50   \n",
       "2                                                     4.75   4.75   4.75   \n",
       "3     0% 16 oz can. Funny story: As I finally wal...  4.75   4.75   4.50   \n",
       "4     Classic TH NEIPA. Overflowing head and bouq...  4.25   4.50   4.25   \n",
       "\n",
       "   feel  overall  score  \n",
       "0  4.25     4.00   4.03  \n",
       "1  4.50     4.50   4.50  \n",
       "2  4.75     4.75   4.75  \n",
       "3  4.50     4.50   4.58  \n",
       "4  4.25     4.25   4.31  "
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = pd.read_csv('C:/Users/HOME/Desktop/8차/통계학 세미나/팀플/잠재요인 돌릴때 쓰는 데이터/reviews.csv')\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-11-27T16:42:31.199539Z",
     "start_time": "2019-11-27T16:42:31.197539Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(9073128, 10)\n"
     ]
    }
   ],
   "source": [
    "print(df.shape) #원본 데이터 900만개를 사용"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-11-27T16:42:35.466500Z",
     "start_time": "2019-11-27T16:42:35.277458Z"
    }
   },
   "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>beerid</th>\n",
       "      <th>userid</th>\n",
       "      <th>rating</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>271781</td>\n",
       "      <td>bluejacket74</td>\n",
       "      <td>4.03</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>125646</td>\n",
       "      <td>_dirty_</td>\n",
       "      <td>4.50</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>125646</td>\n",
       "      <td>CJDUBYA</td>\n",
       "      <td>4.75</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>125646</td>\n",
       "      <td>GratefulBeerGuy</td>\n",
       "      <td>4.58</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>125646</td>\n",
       "      <td>LukeGude</td>\n",
       "      <td>4.31</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   beerid           userid  rating\n",
       "0  271781     bluejacket74    4.03\n",
       "1  125646          _dirty_    4.50\n",
       "2  125646          CJDUBYA    4.75\n",
       "3  125646  GratefulBeerGuy    4.58\n",
       "4  125646         LukeGude    4.31"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#잠깐 전처리\n",
    "\n",
    "cols = ['beer_id','username','score']\n",
    "df = df[cols]\n",
    "df.rename(columns = {'beer_id':'beerid','username':'userid','score':'rating'}, inplace = True)\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-11-27T16:42:43.052209Z",
     "start_time": "2019-11-27T16:42:38.373155Z"
    }
   },
   "outputs": [],
   "source": [
    "#딥러닝에 이용할 컬럼 변경: user, beer, rating\n",
    "\n",
    "user_enc = LabelEncoder()\n",
    "uid = df['userid'].values\n",
    "df['user'] = user_enc.fit_transform(uid.tolist()) #array 인식이 안돼서 list로 빼냄.(userid의 경우만 오류 발생)\n",
    "n_users = df['user'].nunique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(164935, 309542, 1.0, 5.0)"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "item_enc = LabelEncoder()\n",
    "df['beer'] = item_enc.fit_transform(df['beerid'].values)\n",
    "n_beers = df['beer'].nunique()\n",
    "\n",
    "df['rating'] = df['rating'].values.astype(np.float32)\n",
    "min_rating = min(df['rating'])\n",
    "max_rating = max(df['rating'])\n",
    "\n",
    "n_users, n_beers, min_rating, max_rating #n_users, n_beers: 유니크한 값 개수, rating의 최솟값과 최댓값."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-11-27T16:42:46.866068Z",
     "start_time": "2019-11-27T16:42:46.855069Z"
    }
   },
   "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>user</th>\n",
       "      <th>beer</th>\n",
       "      <th>rating</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>102213</td>\n",
       "      <td>231507</td>\n",
       "      <td>4.03</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>94429</td>\n",
       "      <td>112229</td>\n",
       "      <td>4.50</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>15635</td>\n",
       "      <td>112229</td>\n",
       "      <td>4.75</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>34826</td>\n",
       "      <td>112229</td>\n",
       "      <td>4.58</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>55007</td>\n",
       "      <td>112229</td>\n",
       "      <td>4.31</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     user    beer  rating\n",
       "0  102213  231507    4.03\n",
       "1   94429  112229    4.50\n",
       "2   15635  112229    4.75\n",
       "3   34826  112229    4.58\n",
       "4   55007  112229    4.31"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df2=df[['user','beer','rating']] \n",
    "df2.head() #완료된 데이터셋. 여기서 'user','beer','rating'만 사용"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-11-27T16:42:50.351854Z",
     "start_time": "2019-11-27T16:42:49.645695Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((7258502, 2), (1814626, 2), (7258502,), (1814626,))"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#train, test 셋 분할: test size는 0.2, random_state는 임의로 지정\n",
    "X = df[['user', 'beer']].values\n",
    "y = df['rating'].values\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)\n",
    "\n",
    "X_train.shape, X_test.shape, y_train.shape, y_test.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-11-27T16:42:54.302743Z",
     "start_time": "2019-11-27T16:42:54.300743Z"
    }
   },
   "outputs": [],
   "source": [
    "#딥러닝 모델을 만들 때 필요한 또다른 변수는: 잠재요인의 개수를 정하는 것!\n",
    "#잠재요인의 개수는 원하는 대로 설정하면 되지만, 주의할 것은 user와 beer 모두에 대해 같은 사이즈를 줘야 한다는 점.\n",
    "\n",
    "#우리는 50개로 간다!(X에 대해서만 적용)\n",
    "n_factors = 50\n",
    "\n",
    "#X를 array로 만들기\n",
    "X_train_array = [X_train[:, 0], X_train[:, 1]]\n",
    "X_test_array = [X_test[:, 0], X_test[:, 1]]"
   ]
  },
  {
   "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>id</th>\n",
       "      <th>name</th>\n",
       "      <th>brewery_id</th>\n",
       "      <th>state</th>\n",
       "      <th>country</th>\n",
       "      <th>style</th>\n",
       "      <th>availability</th>\n",
       "      <th>abv</th>\n",
       "      <th>notes</th>\n",
       "      <th>retired</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>202522</td>\n",
       "      <td>Olde Cogitator</td>\n",
       "      <td>2199</td>\n",
       "      <td>CA</td>\n",
       "      <td>US</td>\n",
       "      <td>English Oatmeal Stout</td>\n",
       "      <td>Rotating</td>\n",
       "      <td>7.3</td>\n",
       "      <td>No notes at this time.</td>\n",
       "      <td>f</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>82352</td>\n",
       "      <td>Konrads Stout Russian Imperial Stout</td>\n",
       "      <td>18604</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NO</td>\n",
       "      <td>Russian Imperial Stout</td>\n",
       "      <td>Rotating</td>\n",
       "      <td>10.4</td>\n",
       "      <td>No notes at this time.</td>\n",
       "      <td>f</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>214879</td>\n",
       "      <td>Scottish Right</td>\n",
       "      <td>44306</td>\n",
       "      <td>IN</td>\n",
       "      <td>US</td>\n",
       "      <td>Scottish Ale</td>\n",
       "      <td>Year-round</td>\n",
       "      <td>4.0</td>\n",
       "      <td>No notes at this time.</td>\n",
       "      <td>t</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>320009</td>\n",
       "      <td>MegaMeow Imperial Stout</td>\n",
       "      <td>4378</td>\n",
       "      <td>WA</td>\n",
       "      <td>US</td>\n",
       "      <td>American Imperial Stout</td>\n",
       "      <td>Winter</td>\n",
       "      <td>8.7</td>\n",
       "      <td>Every time this year</td>\n",
       "      <td>f</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>246438</td>\n",
       "      <td>Peaches-N-Cream</td>\n",
       "      <td>44617</td>\n",
       "      <td>PA</td>\n",
       "      <td>US</td>\n",
       "      <td>American Cream Ale</td>\n",
       "      <td>Rotating</td>\n",
       "      <td>5.1</td>\n",
       "      <td>No notes at this time.</td>\n",
       "      <td>f</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       id                                  name  brewery_id state country  \\\n",
       "0  202522                        Olde Cogitator        2199    CA      US   \n",
       "1   82352  Konrads Stout Russian Imperial Stout       18604   NaN      NO   \n",
       "2  214879                        Scottish Right       44306    IN      US   \n",
       "3  320009               MegaMeow Imperial Stout        4378    WA      US   \n",
       "4  246438                       Peaches-N-Cream       44617    PA      US   \n",
       "\n",
       "                     style availability   abv                   notes retired  \n",
       "0    English Oatmeal Stout     Rotating   7.3  No notes at this time.       f  \n",
       "1   Russian Imperial Stout     Rotating  10.4  No notes at this time.       f  \n",
       "2             Scottish Ale   Year-round   4.0  No notes at this time.       t  \n",
       "3  American Imperial Stout       Winter   8.7    Every time this year       f  \n",
       "4       American Cream Ale     Rotating   5.1  No notes at this time.       f  "
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "beers = pd.read_csv('C:/Users/HOME/Desktop/8차/통계학 세미나/팀플/잠재요인 돌릴때 쓰는 데이터/beers.csv')\n",
    "beers.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "## create a dictionary out of beerid, beer original title\n",
    "beers_dict = beers.set_index('id')['name'].to_dict()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "##시각화(각 epoch마다의 MSE, MAE 값 비교)\n",
    "\n",
    "def plot_history(hist):\n",
    "    \n",
    "    plt.figure(figsize=(10,15))\n",
    "\n",
    "    plt.subplot(2,1,1)\n",
    "    plt.xlabel('Epoch')\n",
    "    plt.ylabel('Mean Abs Error [rating]')\n",
    "    plt.plot(hist['epoch'], hist['mae'], label='Train Error')\n",
    "    plt.plot(hist['epoch'], hist['val_mae'],label = 'Test Error')\n",
    "    plt.legend()\n",
    "\n",
    "    plt.subplot(2,1,2)\n",
    "    plt.xlabel('Epoch')\n",
    "    plt.ylabel('Mean Square Error [$rating^2$]')\n",
    "    plt.plot(hist['epoch'], hist['mse'], label='Train Error')\n",
    "    plt.plot(hist['epoch'], hist['val_mse'],label = 'Test Error')\n",
    "    plt.legend()\n",
    "    plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-11-27T16:43:14.561307Z",
     "start_time": "2019-11-27T16:43:05.998377Z"
    }
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "#keras 라이브러리 로드\n",
    "from keras.models import Model\n",
    "from keras.layers import Input, Reshape, Dot\n",
    "from keras.layers import Embedding, Dense, Dropout\n",
    "from keras.optimizers import Adam\n",
    "from keras.regularizers import l2\n",
    "from keras.layers import Add, Activation, Lambda, Concatenate"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 레이어 1개만 추가한 버전, 액티베이션은 relu; test MSE: 0.1816, 과적합"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "##bias가 있음, 아주 단순한 버전. 그냥 R_hat = u*m으로 연산한다. 최저 rating과 최고 rating을 이용한다.\n",
    "##layer 1개를 추가하였고, activation 함수 1개는: relu\n",
    "\n",
    "class EmbeddingLayer:\n",
    "    def __init__(self, n_items, n_factors):\n",
    "        self.n_items = n_items\n",
    "        self.n_factors = n_factors\n",
    "    \n",
    "    def __call__(self, x):\n",
    "        x = Embedding(self.n_items, self.n_factors, embeddings_initializer='he_normal')(x) \n",
    "        x = Reshape((self.n_factors,))(x)\n",
    "        return x\n",
    "\n",
    "def RecommenderV7(n_users, n_beers, n_factors, min_rating, max_rating):\n",
    "    \n",
    "    #input: user, beer\n",
    "    user = Input(shape=(1,))\n",
    "    u = EmbeddingLayer(n_users, n_factors)(user) #u는 user * n_factor\n",
    "    ub = EmbeddingLayer(n_users, 1)(user) #user_bias\n",
    "    \n",
    "    beer = Input(shape=(1,))\n",
    "    b = EmbeddingLayer(n_beers, n_factors)(beer) #b은 beer * n_factor\n",
    "    bb = EmbeddingLayer(n_beers, 1)(beer) #beer_bias\n",
    "    \n",
    "    #input 연산- activation 1가지를 추가. relu\n",
    "    x = Concatenate()([u, b])\n",
    "    \n",
    "    x = Dense(1, kernel_initializer='he_normal')(x)\n",
    "    x = Activation('relu')(x)\n",
    "    \n",
    "    x = Lambda(lambda x: x * (max_rating - min_rating) + min_rating)(x)\n",
    "\n",
    "\n",
    "    #최종 모델 만들기(input은 user, beer 2개를 받고, output은 x를 반환.)\n",
    "    model = Model(inputs=[user, beer], outputs=x)\n",
    "    opt = Adam(lr=0.001) #learning rate(학습률) 설정. #optimizer인 최적화 방법은 Adam을 씀.\n",
    "    model.compile(loss=['mse'], metrics = ['mae','mse'] ,optimizer=opt) #loss와 함께 출력할 mae, mse는 metrics로 설정.\n",
    "\n",
    "    return model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"model_1\"\n",
      "__________________________________________________________________________________________________\n",
      "Layer (type)                    Output Shape         Param #     Connected to                     \n",
      "==================================================================================================\n",
      "input_1 (InputLayer)            (None, 1)            0                                            \n",
      "__________________________________________________________________________________________________\n",
      "input_2 (InputLayer)            (None, 1)            0                                            \n",
      "__________________________________________________________________________________________________\n",
      "embedding_1 (Embedding)         (None, 1, 50)        8246750     input_1[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "embedding_3 (Embedding)         (None, 1, 50)        15477100    input_2[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "reshape_1 (Reshape)             (None, 50)           0           embedding_1[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "reshape_3 (Reshape)             (None, 50)           0           embedding_3[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "concatenate_1 (Concatenate)     (None, 100)          0           reshape_1[0][0]                  \n",
      "                                                                 reshape_3[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "dense_1 (Dense)                 (None, 1)            101         concatenate_1[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "activation_1 (Activation)       (None, 1)            0           dense_1[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "lambda_1 (Lambda)               (None, 1)            0           activation_1[0][0]               \n",
      "==================================================================================================\n",
      "Total params: 23,723,951\n",
      "Trainable params: 23,723,951\n",
      "Non-trainable params: 0\n",
      "__________________________________________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "re7model = RecommenderV7(n_users, n_beers, n_factors, min_rating, max_rating)\n",
    "re7model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From C:\\Users\\HOME\\Anaconda3\\lib\\site-packages\\keras\\backend\\tensorflow_backend.py:422: The name tf.global_variables is deprecated. Please use tf.compat.v1.global_variables instead.\n",
      "\n",
      "Train on 7258502 samples, validate on 1814626 samples\n",
      "Epoch 1/10\n",
      "7258502/7258502 [==============================] - 2472s 341us/step - loss: 0.3126 - mae: 0.3788 - mse: 0.3126 - val_loss: 0.1896 - val_mae: 0.3093 - val_mse: 0.1896\n",
      "Epoch 2/10\n",
      "7258502/7258502 [==============================] - 2460s 339us/step - loss: 0.1814 - mae: 0.3025 - mse: 0.1814 - val_loss: 0.1850 - val_mae: 0.3056 - val_mse: 0.1850\n",
      "Epoch 3/10\n",
      "7258502/7258502 [==============================] - 2469s 340us/step - loss: 0.1753 - mae: 0.2960 - mse: 0.1753 - val_loss: 0.1828 - val_mae: 0.3030 - val_mse: 0.1828\n",
      "Epoch 4/10\n",
      "7258502/7258502 [==============================] - 2470s 340us/step - loss: 0.1721 - mae: 0.2925 - mse: 0.1721 - val_loss: 0.1828 - val_mae: 0.3030 - val_mse: 0.1828\n",
      "Epoch 5/10\n",
      "7258502/7258502 [==============================] - 2474s 341us/step - loss: 0.1700 - mae: 0.2904 - mse: 0.1700 - val_loss: 0.1826 - val_mae: 0.3012 - val_mse: 0.1826\n",
      "Epoch 6/10\n",
      "7258502/7258502 [==============================] - 2475s 341us/step - loss: 0.1686 - mae: 0.2889 - mse: 0.1686 - val_loss: 0.1823 - val_mae: 0.3016 - val_mse: 0.1823\n",
      "Epoch 7/10\n",
      "7258502/7258502 [==============================] - 2475s 341us/step - loss: 0.1675 - mae: 0.2877 - mse: 0.1675 - val_loss: 0.1815 - val_mae: 0.3010 - val_mse: 0.1815\n",
      "Epoch 8/10\n",
      "7258502/7258502 [==============================] - 2466s 340us/step - loss: 0.1667 - mae: 0.2868 - mse: 0.1667 - val_loss: 0.1821 - val_mae: 0.2998 - val_mse: 0.1821\n",
      "Epoch 9/10\n",
      "7258502/7258502 [==============================] - 2466s 340us/step - loss: 0.1660 - mae: 0.2861 - mse: 0.1660 - val_loss: 0.1814 - val_mae: 0.3000 - val_mse: 0.1814\n",
      "Epoch 10/10\n",
      "7258502/7258502 [==============================] - 2471s 340us/step - loss: 0.1653 - mae: 0.2854 - mse: 0.1653 - val_loss: 0.1817 - val_mae: 0.3014 - val_mse: 0.1817\n"
     ]
    }
   ],
   "source": [
    "## fit model\n",
    "seed = 7\n",
    "np.random.seed(seed)\n",
    "history_mf = re7model.fit(x=X_train_array, y=y_train, batch_size=1000, epochs=10,\n",
    "                    verbose=1, validation_data=(X_test_array, y_test)) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "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>val_loss</th>\n",
       "      <th>val_mae</th>\n",
       "      <th>val_mse</th>\n",
       "      <th>loss</th>\n",
       "      <th>mae</th>\n",
       "      <th>mse</th>\n",
       "      <th>epoch</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.189635</td>\n",
       "      <td>0.309336</td>\n",
       "      <td>0.189635</td>\n",
       "      <td>0.312647</td>\n",
       "      <td>0.378812</td>\n",
       "      <td>0.312648</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.185034</td>\n",
       "      <td>0.305646</td>\n",
       "      <td>0.185034</td>\n",
       "      <td>0.181390</td>\n",
       "      <td>0.302504</td>\n",
       "      <td>0.181391</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.182810</td>\n",
       "      <td>0.302957</td>\n",
       "      <td>0.182809</td>\n",
       "      <td>0.175254</td>\n",
       "      <td>0.295971</td>\n",
       "      <td>0.175254</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.182779</td>\n",
       "      <td>0.302990</td>\n",
       "      <td>0.182779</td>\n",
       "      <td>0.172056</td>\n",
       "      <td>0.292497</td>\n",
       "      <td>0.172057</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.182570</td>\n",
       "      <td>0.301165</td>\n",
       "      <td>0.182570</td>\n",
       "      <td>0.170035</td>\n",
       "      <td>0.290379</td>\n",
       "      <td>0.170035</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.182333</td>\n",
       "      <td>0.301578</td>\n",
       "      <td>0.182333</td>\n",
       "      <td>0.168620</td>\n",
       "      <td>0.288869</td>\n",
       "      <td>0.168620</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0.181471</td>\n",
       "      <td>0.301017</td>\n",
       "      <td>0.181471</td>\n",
       "      <td>0.167543</td>\n",
       "      <td>0.287732</td>\n",
       "      <td>0.167543</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0.182095</td>\n",
       "      <td>0.299840</td>\n",
       "      <td>0.182095</td>\n",
       "      <td>0.166696</td>\n",
       "      <td>0.286813</td>\n",
       "      <td>0.166696</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>0.181361</td>\n",
       "      <td>0.299953</td>\n",
       "      <td>0.181361</td>\n",
       "      <td>0.165972</td>\n",
       "      <td>0.286074</td>\n",
       "      <td>0.165972</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>0.181679</td>\n",
       "      <td>0.301357</td>\n",
       "      <td>0.181679</td>\n",
       "      <td>0.165346</td>\n",
       "      <td>0.285448</td>\n",
       "      <td>0.165346</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   val_loss   val_mae   val_mse      loss       mae       mse  epoch\n",
       "0  0.189635  0.309336  0.189635  0.312647  0.378812  0.312648      0\n",
       "1  0.185034  0.305646  0.185034  0.181390  0.302504  0.181391      1\n",
       "2  0.182810  0.302957  0.182809  0.175254  0.295971  0.175254      2\n",
       "3  0.182779  0.302990  0.182779  0.172056  0.292497  0.172057      3\n",
       "4  0.182570  0.301165  0.182570  0.170035  0.290379  0.170035      4\n",
       "5  0.182333  0.301578  0.182333  0.168620  0.288869  0.168620      5\n",
       "6  0.181471  0.301017  0.181471  0.167543  0.287732  0.167543      6\n",
       "7  0.182095  0.299840  0.182095  0.166696  0.286813  0.166696      7\n",
       "8  0.181361  0.299953  0.181361  0.165972  0.286074  0.165972      8\n",
       "9  0.181679  0.301357  0.181679  0.165346  0.285448  0.165346      9"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "## show loss & mae & mse at each epoch\n",
    "hist7 = pd.DataFrame(history_mf.history)\n",
    "hist7['epoch'] = history_mf.epoch\n",
    "hist7"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1814626/1814626 [==============================] - 4s 2us/step\n",
      "## evaluation loss and_metrics ##\n",
      "[0.181679046346107, 0.30135658383369446, 0.18167895078659058]\n"
     ]
    }
   ],
   "source": [
    "#모델 평가하기; model.evaluate()\n",
    "loss_and_metrics = re7model.evaluate(X_test_array, y_test, batch_size=1000)\n",
    "print('## evaluation loss and_metrics ##')\n",
    "print(loss_and_metrics)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAm0AAANcCAYAAAANWpgCAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjAsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+17YcXAAAgAElEQVR4nOzdeXycZ333++9vNu2akWV503jP5kUjxzjBDimBECBAIVCgQAihUF558hxSKMtpU5pTQkqfk3LaAIUcKOVJShJIoCxtCqShaSFAcBYntqXYJoljYlu2Y8uyrX0dXc8f90gayZIsaWZ0z0if9+t1v2bubeYnK4m/ua77ui5zzgkAAAD5LeB3AQAAADg3QhsAAEABILQBAAAUAEIbAABAASC0AQAAFICQ3wXk2sKFC92qVav8LgMAAOCcnn766ZPOuZrxzs350LZq1Srt2LHD7zIAAADOycwOTnSO7lEAAIACQGgDAAAoAIQ2AACAAjDnn2kDAADT19/fr6amJvX09PhdypxUXFyseDyucDg85XsIbQAA4CxNTU2qqKjQqlWrZGZ+lzOnOOfU0tKipqYmrV69esr30T0KAADO0tPTo+rqagJbDpiZqqurp92KSWgDAADjIrDlzkz+bAltAAAABcCX0GZmV5vZc2a238xuHuf8jWbWaGa7zOzXZrY+dTxsZt9KndtnZn8x+9UDAIBca2lp0aZNm7Rp0yYtWbJEtbW1w/t9fX1T+owPfehDeu6556b8nd/85jdVU1Mz/D2bNm2a1v25NusDEcwsKOlOSa+X1CTpKTN70Dm3N+2y7zjnvp66/m2S7pB0taR3SypyztWZWamkvWZ2v3PupVn9IQAAQE5VV1dr165dkqRbb71V5eXl+vSnPz3qGuecnHMKBMZvg7r77run/b3vf//79aUvfWnC8wMDAwqFRuLTuWpIl0wmFQwGp13TED9a2i6VtN85d8A51yfpAUnXpF/gnGtL2y2T5IZOSSozs5CkEkl9ktKvBQAAc9j+/fu1ceNG3Xjjjdq8ebOOHTumG264QVu2bNGGDRt02223DV97+eWXa9euXRoYGFAsFtPNN9+s+vp6bdu2TSdOnJjydz7yyCO66qqr9N73vlcXX3zxuDXcd999qqur08aNG/WZz3xGkoa/95ZbbtGll16qJ598MqOf3Y8pP2olHU7bb5L0yrEXmdlHJX1SUkTSlanD35cX8I5JKpX0CefcqXHuvUHSDZK0YsWKbNYOAMC887l/36O9R7PbRrJ+WaU++9YNM7p37969uvvuu/X1r39dknT77bdrwYIFGhgY0Gtf+1q9613v0vr160fd09raqiuuuEK33367PvnJT+quu+7SzTef9YSWvv3tb+sXv/jF8P5Q0Hr88ce1d+9erVixQvv37x9VQ1NTk2655Rbt2LFD0WhUV111lX784x/r6quvVmtrqzZv3qzPf/7zM/pZ0/nR0jbecAl31gHn7nTOrZX055JuSR2+VFJS0jJJqyV9yszWjHPvN5xzW5xzW2pqarJXOQAA8N3atWt1ySWXDO/ff//92rx5szZv3qx9+/Zp7969Z91TUlKiN73pTZKkV7ziFXrppZfG/ez3v//92rVr1/AWiUQkSdu2bRvVEJRewxNPPKErr7xSCxcuVDgc1rXXXqtf/vKXkqRIJKJ3vOMdWfm5/Whpa5K0PG0/LunoJNc/IOlrqffXSvoP51y/pBNm9pikLZIO5KJQAACgGbeI5UpZWdnw+xdeeEFf/vKX9eSTTyoWi+m6664bd/6zofAlScFgUAMDAzP+zrH7zp3V9jSspKQka1On+NHS9pSk881stZlFJL1X0oPpF5jZ+Wm7b5H0Qur9IUlXmqdM0lZJv52FmifU3N6rLz3yvF443u5nGQAAzEttbW2qqKhQZWWljh07pocffnjWa9i6dat+/vOfq6WlRQMDA3rggQd0xRVXZP17Zr2lzTk3YGY3SXpYUlDSXc65PWZ2m6QdzrkHJd1kZldJ6pd0WtIHU7ffKeluSc/K62a92znXMNs/Q7rkoNOXHnlB0ZKwzl9c4WcpAADMO5s3b9b69eu1ceNGrVmzRq961asy+ryxz7T94z/+4znvicfjuu222/Sa17xGzjm99a1v1Vve8pZpt+adi03WpDcXbNmyxe3YsSNnn++c06X/6790+XkL9cX3bMrZ9wAAMJv27dundevW+V3GnDben7GZPe2c2zLe9ayIkCEzU308qoamM36XAgAA5jBCWxbU1cZ04GSn2nv6/S4FAADMUYS2LEgsj8o56dkjzPMLAAByg9CWBXW1UUlS4xG6SAEAQG4Q2rJgYXmRamMl2t3U6ncpAABgjiK0ZUkiHlUjoQ0AAOSIHysizEl18ageevZlnenqU6w0cu4bAADAhFpaWvS6171OkvTyyy8rGAxqaGnKJ598ctQKB5O566679OY3v1lLliw569x1112nxx57TNGo95hTRUWFfvWrX2XpJ8g+QluW1MdjkqSGpla9+gLWOwUAIBPV1dXatWuXJOnWW29VeXm5Pv3pT0/7c+666y5t3rx53NAmSV/84hf19re/fcL7BwYGFAqFJtyf6n3ZQGjLko3LhgYjENoAAMilb33rW7rzzjvV19enyy67TF/96lc1ODioD33oQ9q1a5ecc7rhhhu0ePFi7dq1S+95z3tUUlIy5Ra6W265Rc3NzTpw4ICWLFmiK664Qo888og6OjrU29urhx9+WJ/+9Kf1s5/9TGamz372s3rXu96lRx55RLfffrsWLlyoPXv2qLGxMas/N6EtS6KlYa2qLtXuw4wgBQDMMQ/dLL2c3QCiJXXSm26f9m3PPvusfvSjH+k3v/mNQqGQbrjhBj3wwANau3atTp48ORyUzpw5o1gspq985Sv66le/qk2bxl+16BOf+IRuvfVWSVIikdA999wjSdq5c6d++ctfqri4WN/85je1fft27dq1S1VVVfrud7+rvXv3avfu3WpubtYll1yiV7/61ZKkxx9/XHv37tWKFStm8IcyOUJbFiXiMT310im/ywAAYM565JFH9NRTT2nLFm+lp+7ubi1fvlxvfOMb9dxzz+njH/+43vzmN+sNb3jDlD5vou7Ra665RsXFxcP7b3jDG1RVVSVJ+vWvf61rr71WwWBQS5Ys0eWXX64dO3YoEolo27ZtOQlsEqEtqxLxqB7cfVQn2nu0qKL43DcAAFAIZtAilivOOX34wx/WX//1X591rqGhQQ899JD+4R/+QT/4wQ/0jW98Y8bfU1ZWNuH+ZOu2j70vm5jyI4sSqcEITP0BAEBuXHXVVfre976nkydPSvJGmR46dEjNzc1yzund7363Pve5z+mZZ56R5I0IbW9vz2oNr371q/XAAw8omUzq+PHjeuyxx4Zb/nKJlrYs2rCsUmbeCNLXrVvsdzkAAMw5dXV1+uxnP6urrrpKg4ODCofD+vrXv65gMKg//uM/lnNOZqa//du/lSR96EMf0kc+8pEJByKkP9MmSU8//fQ5a3jXu96lxx9/XPX19TIz3XHHHVq0aFFWf87x2GRNfHPBli1b3I4dO2bt+15/x6OKV5Xo7g9dOmvfCQBAtu3bt0/r1q3zu4w5bbw/YzN72jk3brMd3aNZlojH1HikddL+bgAAgOkitGVZIh7VyY4+HWvt8bsUAAAwhxDasiwR9ybZbWhivjYAQGGj1yh3ZvJnS2jLsnVLKxUKmBoYQQoAKGDFxcVqaWkhuOWAc04tLS2j5oGbCkaPZllxOKgLFlcQ2gAABS0ej6upqUnNzc1+lzInFRcXKx6PT+seQlsO1C+P6icNx4aHHQMAUGjC4bBWr17tdxlIQ/doDtTVxtTWM6BDp7r8LgUAAMwRhLYcGBqMsJsuUgAAkCWEthy4YHGFIqGAGhlBCgAAsoTQlgORUEDrllbS0gYAALKG0JYj9fGo9hxpVXKQodIAACBzhLYcqauNqrMvqd+d7PC7FAAAMAcQ2nKkfnlMkrT7MF2kAAAgc4S2HFlbU66ScFCNRwhtAAAgc4S2HAkGTBtrK1mDFAAAZAWhLYcS8Zj2HG1Tf3LQ71IAAECBI7TlUCIeVe/AoF44zmAEAACQGUJbDiXi3mAEukgBAECmCG05tHJBqSqKQ2pgMAIAAMgQoS2HAgFTXW1UjayMAAAAMkRoy7FEPKbfvtym3oGk36UAAIACRmjLsUQ8qv6k02+PtftdCgAAKGCEthxLxKOSGIwAAAAyQ2jLsdpYiRaURdTAc20AACADhLYcM0sNRmAEKQAAyAChbRbUx6N6/ni7uvoG/C4FAAAUKELbLKiLxzTopL1H2/wuBQAAFChC2ywYGoywm+faAADADBHaZsHiymItrixSIyNIAQDADBHaZkldbYzlrAAAwIwR2mZJfTyqA82dauvp97sUAABQgAhts6Qu9Vzbs7S2AQCAGSC0zZJEPCZJTLILAABmhNA2SxaURRSvKlEjoQ0AAMwAoW0WJeJRNRxhBCkAAJg+QtssSsRjOnyqW6c6+/wuBQAAFBhC2yxK1HqDEViHFAAATBehbRZtTI0gbThMFykAAJgeQtssqiwOa83CMibZBQAA00Zom2V18SgjSAEAwLQR2mZZIh7Ty209OtHW43cpAACggBDaZlli6Lk2WtsAAMA0ENpm2YZllQqY1NDEYAQAADB1hLZZVhoJ6fxFFQxGAAAA00Jo88HQYATnnN+lAACAAkFo80F9PKqWzj4dOdPtdykAAKBAENp8UBePSRJTfwAAgCkjtPlg3dIKhYOm3YQ2AAAwRb6ENjO72syeM7P9ZnbzOOdvNLNGM9tlZr82s/Vp5xJmtt3M9qSuKZ7d6jNXFArqwiUVajzCCFIAADA1sx7azCwo6U5Jb5K0XtL70kNZynecc3XOuU2SviDpjtS9IUn3SbrRObdB0msk9c9W7dlUVxtTA4MRAADAFPnR0nappP3OuQPOuT5JD0i6Jv0C51xb2m6ZpKFk8wZJDc653anrWpxzyVmoOevq41G19wzopZYuv0sBAAAFwI/QVivpcNp+U+rYKGb2UTN7UV5L28dShy+Q5MzsYTN7xsz+bLwvMLMbzGyHme1obm7OcvnZUTe8MgJdpAAA4Nz8CG02zrGz+gidc3c659ZK+nNJt6QOhyRdLun9qdd3mNnrxrn3G865Lc65LTU1NdmrPIsuWFyholCAEaQAAGBK/AhtTZKWp+3HJR2d5PoHJL097d5HnXMnnXNdkn4qaXNOqsyxcDCg9csqWYMUAABMiR+h7SlJ55vZajOLSHqvpAfTLzCz89N23yLphdT7hyUlzKw0NSjhCkl7Z6HmnEjURvXs0VYlBxmMAAAAJjfroc05NyDpJnkBbJ+k7znn9pjZbWb2ttRlN6Wm9Ngl6ZOSPpi697S8kaRPSdol6Rnn3E9m+2fIlkQ8pq6+pF5s7vC7FAAAkOdCfnypc+6n8ro204/9Vdr7j09y733ypv0oeInhwQitumBxhc/VAACAfMaKCD5aU1OuskhQjYwgBQAA50Bo81EwYNpQG2U5KwAAcE6ENp8laqPae6xN/clBv0sBAAB5jNDms8TymPoGBvXcy+1+lwIAAPIYoc1niVpvMELjEbpIAQDAxAhtPltZXarK4hCT7AIAgEkR2nxmZkrEY6xBCgAAJkVoywN18aiee7ldPf1Jv0sBAAB5itCWB+rjUQ0MOu071uZ3KQAAIE8R2vJAXTwmicEIAABgYoS2PLAsWqzqsgiDEQAAwIQIbXnAG4wQZTACAACYEKEtT9TFY9p/okOdvQN+lwIAAPIQoS1P1MejGnTSnqMMRgAAAGcjtOWJutTKCHSRAgCA8RDa8sSiymItqSxmBCkAABgXoS2PeIMRCG0AAOBshLY8kohH9buTnWrt7ve7FAAAkGcIbXkkkZpk91m6SAEAwBiEtjwyMhiB0AYAAEYjtOWRqrKIli8oUeMRRpACAIDRCG15JhGPafdhWtoAAMBohLY8k6iN6siZbrV09PpdCgAAyCOEtjwzNBihgcEIAAAgDaEtz2ysrZQkNTIYAQAApCG05ZmK4rDW1JQxghQAAIxCaMtD9fEYa5ACAIBRCG15qK42qhPtvTre1uN3KQAAIE8Q2vJQ/XJvkt3dh2ltAwAAHkJbHlq/NKqASY2MIAUAACmEtjxUEgnqgsUVDEYAAADDCG15KhGPqqHpjJxzfpcCAADyAKEtT9XFYzrd1a+m091+lwIAAPIAoS1P1ce9wQg81wYAACRCW966cEmFwkHTbuZrAwAAIrTlraJQUBctqWQ5KwAAIInQltcS8agam1o1OMhgBAAA5jtCWx5LxKNq7x3QSy2dfpcCAAB8FpruDWb2ySlc1umc+8cZ1IM0iXhMkjcYYU1Nuc/VAAAAP82kpe3/llQuqWKS7VPZKnA+O39RuYpCAe0+zHNtAADMd9NuaZN0r3PutskuMLOyGdaDNKFgQBuWVarxCCNIAQCY76bd0uac+7NsXIOpScRjevZImwaSg36XAgAAfDSTljZJEz7b1irpaefcrpmXhHSJeFT//JuX9GJzpy5cUuF3OQAAwCeZjB7dIulGSbWp7QZJr5H0T2ZGS1uWDA1GaGCSXQAA5rVMQlu1pM3OuU855z4lL8TVSHq1pD/KQm2QtGZhmcoiQTUwyS4AAPNaJqFthaS+tP1+SSudc92SejOqCsMCAdPG2qgaWIMUAIB5bcbPtEn6jqTHzezfUvtvlXR/auTo3owrw7D65TH982MvqW9gUJEQ8yEDADAfzTi0Oef+2swekvQqSSbpRufcjtTp92ejOHjqaqPqSw7q+ePt2lgb9bscAADgg0xa2iRpp6SjQ59jZiucc4cyrgqj1A8PRmgltAEAME/NuK/NzP5E0nFJ/ynpx5J+knpFli1fUKJoSZgRpAAAzGOZtLR9XNKFzrmWbBWD8ZmZEvEoI0gBAJjHMnmq/bC8yXQxCxLxqJ473q6e/qTfpQAAAB9k0tJ2QNIvzOwnSpviwzl3R8ZV4Sx1tTElB532HmvT5hVVfpcDAABmWSYtbYfkPc8WkVSRtiEH6pd7AxAa6SIFAGBeymTKj89lsxBMbkllsRaWF2k3gxEAAJiXph3azOxLzrk/NbN/l+TGnnfOvS0rlWGUocEItLQBADA/zaSl7d7U699lsxCcWyIe1c+fO6GO3gGVF2U6xR4AACgk036mzTn3dOrtJufco+mbpE3ZLQ/pEvGonJP2sA4pAADzTiYDET44zrE/yuDzcA51td7KCI2ENgAA5p2ZPNP2PknXSlptZg+mnaqQxES7OVRTUaRl0WLt5rk2AADmnZk8GPUbScckLZT092nH2yU1ZKMoTKwuHlUjI0gBAJh3ph3anHMHJR2UtC375eBcEvGYHt5zXK1d/YqWhv0uBwAAzJJMFozfamZPmVmHmfWZWdLM2rJZHM6WiKcm2eW5NgAA5pVMBiJ8VdL7JL0gqUTSRyR9ZSo3mtnVZvacme03s5vHOX+jmTWa2S4z+7WZrR9zfkUqLH46g/oLUiI1GKHhCF2kAADMJ5mENjnn9ksKOueSzrm7Jb32XPeYWVDSnZLeJGm9pPeNDWWSvuOcq3PObZL0BUlj1zP9oqSHMqm9UEVLw1pZXaqGw7S0AQAwn2QyQ2uXmUUk7TKzL8gbnFA2hfsulbTfOXdAkszsAUnXSNo7dIFzLr2btUxpKy+Y2dvlLVbfmUHtBa2uNqqdh2hpAwBgPsmkpe0Dqftvkheglkt65xTuq5V0OG2/KXVsFDP7qJm9KK+l7WOpY2WS/lzSpOuemtkNZrbDzHY0NzdPoaTCUh+P6ciZbp3s6PW7FAAAMEtmFNpSXZx/45zrcc61Oec+55z7ZKq79Jy3j3NsvDVM73TOrZUX0m5JHf6cpC865zom+wLn3Decc1ucc1tqamqmUFJhqRsajMB8bQAAzBszCm3OuaSkmlT36HQ1yWuVGxKXdHSS6x+Q9PbU+1dK+oKZvSTpTyV9xsxumkENBW1jbVRmUgOhDQCAeSOTZ9pekvRYalWE4efLnHNjBw2M9ZSk881staQjkt4rb4WFYWZ2vnPuhdTuW+SNUJVz7vfSrrlVUodz7qsZ/AwFqbwopLU15Wpgkl0AAOaNTELb0dQWkLeE1ZQ45wZSrWMPSwpKuss5t8fMbpO0wzn3oKSbzOwqSf2STmv8dU7ntURtVL/af1LOOZmN1+MMAADmkhmHNufcpIMBznHvTyX9dMyxv0p7//EpfMatM/3+uSARj+qHO4/oeFuvlkSL/S4HAADk2LSfaUt1S2Z8DTJTF/cm2d1NFykAAPPCTFraPnKO5apM3nNqt86oIkzJhmWVCgZMjU2teuOGJX6XAwAAcmwmoe2fdO5n2P5pBp+LaSgOB3XB4gpa2gAAmCemHdoyeZYN2ZWojerhvS8zGAEAgHkgo7VH4a/E8qjOdPWr6XS336UAAIAcI7QVsEQtgxEAAJgvZryMlZl9ItvFYHouXFKhSDDAclYAAMwDmSxjdU2Wa8E0RUIBrVvKYAQAAOaDTLpHHzOzr5rZ75nZ5qEta5VhSuriUT17pE2Dg87vUgAAQA5lsozVZanX29KOOUlXZvCZmKZEPKb7Hj+k37V0am1Nud/lAACAHMlkGavXZrMQzEwiHpUkNTSdIbQBADCHzbh71MyiZnaHme1IbX9vZtFsFodzO6+mXMXhgBoYjAAAwJyWyTNtd0lql/SHqa1N0t3ZKApTFwoGtHFZlNAGAMAcl8kzbWudc+9M2/+cme3KtCBMX108qvufPKSB5KBCQabeAwBgLsrkb/huM7t8aMfMXiWJqfl9UB+Pqad/UPubO/wuBQAA5EgmLW03Sron7Tm205I+mHlJmK66ocEIh1t10ZJKn6sBAAC5MNMVEQKSLnTO1UtKSEo45y52zjVktTpMyerqMlUUhdRwhEl2AQCYq2a6IsKgpJtS79ucc21ZrQrTEgiYNtYyGAEAgLksk2fa/tPMPm1my81swdCWtcowLYl4VPuOtal3IOl3KQAAIAcyeabtw6nXj6Ydc5LWZPCZmKFEPKb+pNPzL3cMP+MGAADmjhmFttQzbdc55x7Lcj2YoaGVEXY3nSG0AQAwB2XyTNvfZbkWZCBeVaKq0rAaea4NAIA5KZNn2n5mZu80M8taNZgxM1NdPKbdTYwgBQBgLsoktH1S0r9I6jWzNjNrNzNGkfooURvVCyc61N3HYAQAAOaaGYc251yFcy7gnIs45ypT+8zs6qNEPKrkoNPeY2RnAADmmmmHNjO7Lu39q8acuykbRWFmEvGYJKmBLlIAAOacmbS0fTLt/VfGnPuw4JvFlUWqqShiMAIAAHPQTEKbTfB+vH3MIjNTfTzKYAQAAOagmYQ2N8H78fYxy+pqYzpwslPtPf1+lwIAALJoJpPrXmRmDfJa1dam3iu1z2oIPkssj8o5ac/RNm1dU+13OQAAIEtmEtrWZb0KZE1drbcaQkPTGUIbAABzyLRDm3PuYC4KQXYsLC9SbaxEDQxGAABgTslkcl3kqUQ8qsYjhDYAAOYSQtscVBeP6mBLl8509fldCgAAyJKshDYzqzKzRDY+C5mrT02yS2sbAABzx4xDm5n9wswqzWyBpN2S7jazO7JXGmZq47KhwQiENgAA5opMWtqizrk2SX8g6W7n3CskXZWdspCJaGlYq6pLWc4KAIA5JJPQFjKzpZL+UNKPs1QPsiQRj7GcFQAAc0gmoe02SQ9LetE595SZrZH0QnbKQqYS8aiOtvaoub3X71IAAEAWzDi0Oef+xTmXcM79z9T+AefcO7NXGjKRGB6MQBcpAABzQSYDEdaY2b+bWbOZnTCzfzOz1dksDjO3YVmlzKTdh+kiBQBgLsike/Q7kr4naamkZZL+RdID2SgKmSsrCum8mnKm/QAAYI7IJLSZc+5e59xAartPkstWYchcIh5TQ1OrnOPXAgBAoZt2aDOzBam52X5uZjeb2SozW2lmfybpJ9kvETOViEd1sqNXx1p7/C4FAABkaNoLxkt6Wl6LmqX2/0faOSfprzMtCtmRiI9MsrssVuJzNQAAIBPTDm3OuQkHG5hZOLNykE3rllYqFDA1NJ3R1RuX+F0OAADIQMZrj5rnSjP7pqSmLNSELCkOB3XB4goGIwAAMAdkMuXHK83sy5IOSnpQ0q8kXZStwpAd9cujDEYAAGAOmMlAhL8xsxck/S9JjZIultTsnPuWc+50tgtEZupqY2rt7tehU11+lwIAADIwk5a2GyQdl/Q1Sfc551rEVB95K30wAgAAKFwzCW1LJP2NpLdJ2m9m90oqMbOZjERFjl2wuEKRUEANTSxnBQBAIZvJ6NGkpIckPWRmxZJ+X1KppCNm9l/OuWuzXCMyEAkFtG5pJS1tAAAUuIxGjzrnepxz308tFH++pIezUxayqT4e1bNHWjU4SC82AACFKuMpP4Y459qcc9/K1uche+pqo+rsS+rAyQ6/SwEAADOUtdCG/FW/PCaJwQgAABQyQts8sLamXCXhIKENAIACltGITzO7TNKq9M9xzt2TYU3IsmDAtLG2khGkAAAUsBmHttRUH2sl7ZKUTB12kghteSgRj+m+xw9qIDmoUJAGVgAACk0mLW1bJK13rI9UEBLxqHoHBvX88Q6tX1bpdzkAAGCaMmlyeVbeRLsoAIm4Nxih8QhdpAAAFKJMWtoWStprZk9K6h066Jx7W8ZVIetWLihVRXFIu5ta9Z5L/K4GAABMVyah7dZsFYHcCwRMdbVRNTKCFACAgjTj0OacezSbhSD3EvGY/vevD6h3IKmiUNDvcgAAwDTM+Jk2M9tqZk+ZWYeZ9ZlZ0szapnjv1Wb2nJntN7Obxzl/o5k1mtkuM/u1ma1PHX+9mT2dOve0mV050/rno0Q8qv6k02+PtftdCgAAmKZMBiJ8VdL7JL0gqUTSR1LHJmVmQUl3SnqTpPWS3jcUytJ8xzlX55zbJOkLku5IHT8p6a3OuTpJH5R0bwb1zzuJeFSS1HCELlIAAApNpgvG75cUdM4lnXN3S3rNFG67VNJ+59wB51yfpAckXTPmc9Nb7Mrkzf8m59xO59zR1PE9ktOXjZcAACAASURBVIrNrCiTn2E+qY2VaEFZRA2HGUEKAEChyWQgQpeZRSTtMrMvSDomL2CdS62kw2n7TZJeOfYiM/uopE9Kikgarxv0nZJ2Oud6x54wsxsk3SBJK1asmEJJ84NZajACLW0AABScTFraPpC6/yZJnZKWywtS52LjHDtrgl7n3J3OubWS/lzSLaM+wGyDpL+V9D/G+wLn3Decc1ucc1tqamqmUNL8UR+P6vnj7eruS577YgAAkDcyGT160MxKJC11zn1uGrc2yQt4Q+KSjk5wreR1n35taMfM4pJ+JOl659yL0/heSKqLxzTopD1HW7Vl1QK/ywEAAFOUyejRt8pbd/Q/UvubzOzBKdz6lKTzzWx1qnv1vZJG3Wdm56ftvkXeYAeZWUzSTyT9hXPusZnWPp8ND0ZgvjYAAApKJt2jt8obVHBGkpxzuyStOtdNzrkBeV2qD0vaJ+l7zrk9ZnabmQ2tpnCTme0xs13ynmv74NBxSedJ+n9S04HsMrNFGfwM887iymItriziuTYAAApMJgMRBpxzrWbjPaI2OefcTyX9dMyxv0p7//EJ7vu8pM9P+wsxSl1tTLubGEEKAEAhyWjBeDO7VlLQzM43s69I+k2W6kIO1cejOtDcqfaefr9LAQAAU5RJaPsTSRvkLRZ/v6Q2SX+ajaKQW3Wp59roIgUAoHBkMnq0S9JfpjYUkEQ8JklqbGrVZWsX+lwNAACYimmHtnONEHXOvW2y8/DfgrKI4lUlLGcFAEABmUlL2zZ5KxrcL+kJjT9ZLvJcIh5VA4MRAAAoGDN5pm2JpM9I2ijpy5JeL+mkc+5R59yj2SwOuZOIx3T4VLdOd/b5XQoAAJiCaYe21OLw/+Gc+6CkrZL2S/qFmf1J1qtDziRqU5Ps0kUKAEBBmNHoUTMrMrM/kHSfpI9K+gdJP8xmYcitDanQ1kgXKQAABWEmAxG+Ja9r9CFJn3POPZv1qpBz0ZKw1iwsYzkrAAAKxEwGInxAUqekCyR9LG1FBJPknHOVWaoNOVYXj+qJA6f8LgMAAEzBTJ5pCzjnKlJbZdpWQWArLIl4TC+39ehEW4/fpQAAgHPIZEUEFLhEamUEukgBAMh/hLZ5bMOySgWMEaQAABQCQts8VhoJ6fxFFYwgBQCgABDa5rm6eFQNTa1yzvldCgAAmAShbZ6rj0fV0tmno60MRgAAIJ8R2ua5unhMktRwmC5SAADyGaFtnrtoSYVCAWMwAgAAeY7QNs8Vh4O6aGmFGpn2AwCAvEZog+pqY2poOsNgBAAA8hihDaqPR9XWM6CDLV1+lwIAACZAaIPqUisj7Ga+NgAA8hahDbpgcYWKQgGeawMAII8R2qBwMKD1yyoZQQoAQB4jtEGSlKiN6tkjrUoOMhgBAIB8RGiDJCkRj6mrL6kDzR1+lwIAAMZBaIMkKTE8GIEuUgAA8hGhDZKkNTXlKo0E1cgIUgAA8hKhDZKkYMC0sTbKYAQAAPIUoQ3DErVR7T3apv7koN+lAACAMQhtGJZYHlPvwKCeP97udykAAGAMQhuGJWq9wQgNDEYAACDvENoy1dMqfe1y6Sefkhq/L7Ue8buiGVtZXarK4hChDQCAPBTyu4CC13VKKlso7X5Aeuqb3rHoCmnlNmnFVmnFNmnhhVIg//OxmSkRj6nxCCNIAQDIN4S2TC1YLV3/r1JyQDreKB16XDq0XXrx51LDd71rSqqk5VtTQW6btHSTFIr4W/cE6uJR/dMvD6inP6nicNDvcgAAQAqhLVuCIWnZxd629X9KzkmnDngB7tB26eB26fmHvGtDxVLtlpHWuPilUnGlv/Wn1MejGhh0+u3L7dq0POZ3OQAAIIXQlitmUvVab7v4Ou9Yx4lUiHtcOvgb6Vd/L7lByQLS4o3SystGulQrlvhSdl3cC2qNTWcIbQAA5BFC22wqXyStv8bbJKm3XWp6aiTEPf0t6Ymve+eqVo8OcdXneUEwx5ZFi1VdFtHuplZ9IOffBgAAporQ5qeiCmntld4mScl+6djukda45/9D2vVt71zpQi/ADQW5JfVel2yWeYMRompkBCkAAHmF0JZPgmEpvsXbLvsT77m4ky9Ih34z0hr32x9714bLvOuGQlz8EilSlpUy6uIxPfr8C+rqG1BphH9EAADIB/yNnM/MpJoLvO0Vf+Qdazua9lzcdukXt0tykgWlpfUjIW75Vqm8ZkZfWx+PatBJe4626ZJVC7L24wAAgJkjtBWaymXSxnd6m+RN7nv4yZERqk/+k7T9q9656vNHd6lWrZ7Sc3F1aSsjENoAAMgPhLZCVxyVzn+9t0nSQK90dOdIa9y+B6Wd93rnypeMDnGLN0qBs+diW1RZrCWVxWpoYpJdAADyBaFtrgkVpUacbvX2Bwel5t+mPRe3Xdr7r965SIW0/NKRSX9rXyGFSySJwQgAAOQZQttcFwhIi9d72yUf8Y6dOTx60t///nzq2rA3OfDKbXpbyUr95clytXb3K1oS9q9+AAAgSTLnnN815NSWLVvcjh07/C4jv3Wdkg4/MRLiju6UBvu9U9HzVXre5d5I1UXrpJqLsjZKFQAAjGZmTzvntox7jtCGs/R3q/3Fx/X1e7+j9yxu0oqORqmvPXXSpKqV0qL1XohbtN7bqs/L2/VUAQAoFJOFNrpHcbZwiSoueq0ejDr9rjqq//+mTdLpl6QTe6UT+7zX43ul5x+WXNK7JxDyRqsOB7l13la1atzBDgAAYHoIbZhQIh7T7sNnvNA1tI7qureOXDDQ603+OxTkTuyTjjwt7fnhyDWhEqnmwtEtc4vXSxVLZ2VZLgAA5gpCGyaUqI3qJw3H1NLRq+ryorMvCBVJSzZ6W7reDqn5udEtcy/+t7T7OyPXFEfHdLGmXkuZFw4AgPEQ2jChRDwmSWo80qrXXLho6jcWlUvxV3hbuq5To4PciX3Ssz+Qeu4auaZ88Zggt8FrqSsqz8JPBABA4SK0YUIbayslSY1N0wxtEyldIK263NuGOCe1H0sLc6lAt+NuaaB75LrY2MEP66SF53utfQAAzAOENkyoojisNTVl2p3LSXbNvKW5KpdJ5101cnwwKZ05OHrgw4l90v7/lAYHUvcGvVGr6c/KLVrP4AcAwJxEaMOk6uMx/ebFk7P/xYGgtGCNt130lpHjA31Sy/7RLXPHdkt7/01SavqaUPHZgx8WrZMqaxn8AAAoWIQ2TKquNqof7Tyi4209WlxZ7Hc53lxwQys8pOvrTA1+SHte7sCj0u77R64pqhyZimTRhpFAV1Y9uz8DAAAzQGjDpOqXRyVJDU2tev36PAhtE4mUSbWbvS1d92npxG+lE3tGWub2/Kv09D+PXFO2yGvRCwS9Z+wkSW70e2mC/cnOTXStJr42V98pk8pqpGh8nG25VFpNKyQA5DlCGya1fmlUAZMam87o9esX+13O9JVUSSu3edsQ56SO46NHsp4+6J0byi3DAcbOvT+la8/1OZrGtVP9zrRr3aDUcUI6vsebFDl9kIfkdSmPDXLp7yuXSeESAQD8Q2jDpEoiQa1bWqm7f/OSegYGdd0rV2pFdanfZWXGTKpY4m1rr/S7mtnnnDf9SluT1Dq0HR55v/+/pPaXldYk6CldOHGoi8a9lrxAwJcfCQDmA9YexTk9f7xdX3rkeT2857gGndNrLqjR9dtW6YoLahQI0KU2Jw30Se1Hxw91Q1tfx+h7ghFvsMdZoa52ZD9S5s/PAwAFggXjCW1Z8XJrj77z5CHd/+QhNbf3asWCUl23dYX+cMtyxUpZLH5ecU7qaZ081LUf9bpl05VUjRPq0vbLFzNdC4B5jdBGaMuqvoFBPbznZd27/aCefOmUikIBXbNpma7ftkoba6N+l4d8kRzwJk4eL9i1HfH2e8bMARgIpebtG2ewxND74kp/fp50g0kp2eetv5v+etaxXinZP4Xr+s59fnDAm0w6VCKFi6VwqfcsYrhk5DX9/ahjqXvGuzcY9vtPEygcg8mc/48loY3QljP7jrXpnu0H9a87j6i7P6mLV8R0/baVenPdUhWFaDHBOfS0pQLceK11h6W2oyOTKQ8pio4/CraockzgSQ9MvV6X7zmPTTFYuWT2/gwsIAWLvO7lUMR7H4p4+8GIF9SCRd5fFAO93iCS/p7Ua9r7sX9OU/7+4DjhbiqBb4IQONm9oRIpyKPUyDO9HVLnCamjOfV6QupsTr2OOb7h7dI1d+a0HEIboS3nWrv79YOnm3Tf4wd14GSnqssies8ly/X+rStVG2PUIWZoMOmN9D0r1B0Z2e8+NfXPCxalQlA4LRylh6RzHAuGR0LUhMem+TnZCjHJgbQg1y0N9KS9dqUFvXECX/prf9eYeyf4vLFd31MVCKeFuzGBrzgqlSzwlrwrXZD2vnrkfckC788RmIhzUm/b1EJYZ7P3z/N4Sqq8KaHKF3kDrcoXSSu2ShvekdPy8y60mdnVkr4sKSjpm86528ecv1HSRyUlJXVIusE5tzd17i8k/XHq3Meccw9P9l2Ettk1OOj02Isndc/2g/qvfcclSa9bt1jXb1upy89bKGMuMGRbX+fIwIizWqyKRlqsgmHmossW51KtkzMJfEPHxrm3p9Ub2dx9auK/SCUpUiGVVp0d5kqrU++r0t6nzodL+f0XMue8eTcnC1/pr8necT7EpLKFqSBWc3YgSz9ettC3RwfyKrSZWVDS85JeL6lJ0lOS3jcUylLXVDrn2lLv3ybp/3LOXW1m6yXdL+lSScskPSLpAucm7qsgtPmn6XSXvvPEIT3w1GGd6uzTmoVlum7rSr3zFXFFS3iOBsAk+nu88NbVMhLkulqkrtPjHE9tvZOskxwqHr8Vb1TYG3pf5b0vjhL0cmlw0Pv9TRrCho43S4P9Z3+GBVOh6xwhrHyR97stgIFO+Rbatkm61Tn3xtT+X0iSc+7/neD690m63jn3prHXmtnDqc/aPtH3Edr81zuQ1E8bj+me7Qe189AZlYSDevvFtbp+20qtW5oHD5UDmBuSA15rTFdLWphLe58e8IaCX/fpibt6LThON23VJGFvgVQcm9/P7Q0mvT/XCYPY8ZFjnSfHfz40EJ48fKUfL6mac/NDThba/Pgnq1bS4bT9JkmvHHuRmX1U0iclRSQNzYBaK+nxMffW5qZMZEtRKKh3XBzXOy6Oq7GpVfdsf0k/fKZJ9z95SJesqtIHtq3S1RuWKBKaW//iAZhlwZD3l3t5zdTvGRz0Wui6xga7scHvtHTqd1L3095+sm/izyyOjf9MXqRM3nJ1g2mbG/Oa2s66bpJrR90zwblRx6f6XRN91gTfleyXes6MH4JDxSPhKxqXai+euGWsOEYL5wT8CG3j/SbOau5zzt0p6U4zu1bSLZI+ONV7zewGSTdI0ooVKzIqFtlVF4/q/3t3vT7z5nX6l6cP677HD+lj9+9UTUWR3nfpCl176QotiebxGqcA5pZAwGutKamSqtdO7R7nvGcpR3XTnh6ny7YltWTeb733/Z3eaOH0TTbmmE3wfsyxs+5L3zTJudS9gaBk4Yk/f7zvH/c7beQ1EPIC6qgQlgplRRUEsSwohO7RgKTTzrko3aNzz+Cg06PPN+ue7S/pF883K2CmN25YrA9sXaWtaxYwcAEAMK/kW/foU5LON7PVko5Ieq+ka9MvMLPznXMvpHbfImno/YOSvmNmd8gbiHC+pCdnpWrkRCBgeu1Fi/TaixbpYEunvv3EIX33qcP6aePLumBxuT6wdaXesTmu8qJ5/IwIAADyb8qPN0v6krwpP+5yzv2Nmd0maYdz7kEz+7KkqyT1Szot6Sbn3J7UvX8p6cOSBiT9qXPuocm+i5a2wtPTn9SDu4/q3u0H1XikVeVFIf3BZm/gwnmLKvwuDwCAnMmr0aOzjdBWuJxz2nX4jO7dflA/bjimvuSgtq2p1vXbVur16xcrFGTgAgBgbiG0EdoKXktHr76747C+/fghHTnTraXRYl176Qq999IVqqko8rs8AACygtBGaJszkoNO/7XvuO59/KB+9cJJhYOmN21cquu3rdQrVlYxcAEAUNDybSACMGPBgOkNG5boDRuW6EBzh+59/KC+/3STHtx9VOuWVur6bSt1zaZlKo3wjzYAYG6hpQ0Fr6tvQP+686ju2f6SfvtyuyqKQ3r3K5brA9tWavXCMr/LAwBgyugeJbTNC8457Th4WvdsP6iHGo9pYNDp985fqOu3rdKVFy1SMEDXKQAgvxHaCG3zzon2Hj3w5GF9+4mDOt7Wq9pYid6/dYXes2W5qssZuAAAyE+ENkLbvNWfHNR/7j2ue7a/pMcPnFIkFNDv1y3V9Zet0qblMb/LAwBgFEIboQ2Snj/ernu3H9QPn2lSZ19SiXhUH9i6Um+tX6bicNDv8gAAILQR2pCuvadfP9p5RPdsP6j9JzoUKw3rDy6Oa9vaam1eEaP7FADgG0IboQ3jcM5p+4EW3bv9oP5z73ENDHr/LqysLtXmFVXavCKmi1dU6aIlFay+AACYFczTBozDzHTZ2oW6bO1Cdfcl1XikVc8cOq1nDp7Wr144qR/tPCJJKo0ElYhHU0GuShfTGgcA8AGhDZBUEgnq0tULdOnqBZK8Vrim093DIe6ZQ2f0j788oCStcQAAnxDagHGYmZYvKNXyBaW6ZlOtJKm7L6mGpjN65tAZPXPotH71QvOErXGbV1ZpQVnEzx8BADDHENqAKSqJBPXKNdV65ZpqSV5r3OFT3dp5ePzWuFWp1riLV3otchcupjUOADBzhDZghsxMK6pLtaJ64ta4X77QrB+mtcbVx2PavDKWejaO1jgAwNQR2oAsmqg17plDp4e3rz860hq3emGZLl4eozUOAHBOTPkBzLKuvgE1NA2NVD2jnYdOq6WzTxKtcQAw3zHlB5BHSiMhbV1Tra3TaY1bERse5HDhkgoFA+bnjwAA8AEtbUAemqw1riwSVP3yVIhbGdOm5bTGAcBcQUsbUGDGa407dKprOMQ9c+i0vvboi7TGAcA8QksbUKC6+ga0+7DXGrfz0OStcRcvr1IVrXEAkPdoaQPmoNJISNvWVmvb2qm1xi0sL9LK6lKtTE0avLLa21YsKNPC8ojMaJkDgHxGSxswhw21xu06fEa/O9mhgy1dOnyqS8faepT+r35pJKgVC0q1YsFQkCvViuoyrVxQqtqqEoWZhgQAZgUtbcA8NbY1bkhPf1JNp7t16FSnDrV06eCpLh1q6dLvTnbq0eeb1TswOHxtwKRlsZLhVrn0YLeyulQVxeHZ/rEAYF4itAHzUHE4qPMWleu8ReVnnRscdDrR3quDLZ06dKpLh0516WCL9/rwnpd1KvXc3JCq0vBwq5zXQud1wa6sLtOiiiIFGBABAFlBaAMwSiBgWhIt1pJo8fDKDunae/qHu1kPpgLd4VNd2nn4tH7SeGz4GTpJKgoFvOfnUmFupJWuTMsXlKgoFJzNHw0AChqhDcC0VBSHtbE2qo210bPO9ScHdeR0t9c6d6pLh1KtdQdburT9QIu6+pLD15pJSyqLx32ObmV1qaIlYQZHAEAaQhuArAkHA1q1sEyrFpaddc45p5Mdfaku187hLtdDLV36+XPNam7vHXV9RXEoNdq1bKSVLtVitzRawjx0AOYdQhuAWWFmqqkoUk1FkV6xsuqs8119Azp8qvusZ+n2HmvTz/a+rP7kSLdrOGiKV40e7VobK9GiyiItqihWTUWRisN0vQKYWwhtAPJCaSSkC5dU6MIlFWedSw46HWvtHhnpmmqhO3iqU88cOq32noGz7qksDqmmwgtxXpgrGtmvKNKiyiLVlBersiRENyyAgkBoA5D3ggGvZS1eVarLxpxzzulMV7+OtnbrRHuvmlPbibae4f2dh87oRHuPevoHz/rsolAgFebGCXVp+9XlRXTJAvAVoQ1AQTMzVZVFVFUW0YZJrnPOqb13IBXoenWivWck4LV7+weaO/XE707pTFf/WfcHTFpQVpTWSlc03B07KvBV0jULIDcIbQDmBTNTZXFYlcVhra05e366dL0DyeEwN/ya1nJ3or1X+4616WRH36gpToZUDHfNjgl1Y0IeI2QBTAehDQDGKAoFh7tjJ5McdDrV2ZcKcj2ju2fbe3SirVe7Dk/cNRsJBYZb7CZruasuiyjEUmLAvEdoA4AZCgZGRsSuV+WE1znn1NE74HXDtvWqucN75i69Ne+llk49+dL4XbNmUnVZRLHSiKpKw4qWeK+x0nDqWCT1Pjz8vqo0QjctMMcQ2gAgx8xMFcVhVUyxa/ZkR9/wQIr01rszXX063dWnptNdevZIv850943bgjekOBxQrGRsoBsKdV7gi5WEVVXmvQ6dC9OqB+QlQhsA5JGiUFC1sRLVxkqmdH1Pf1Knu/p0pqt/+HXofWt3v0539ul0V79au/v0wokOnUldMzDOs3hDyotCo1rthlr40oPd2HOVxWHWmQVyjNAGAAWsOBzU0miJlkanFvKkke7a9IB3prvfa8nr9FrwvHNe4Dt8qktnuvvV2t0vN0HWM5OiJWlhLvU+mgp4VaVhRVOvVaUR79qyiMoiQQZjAFNEaAOAeSa9u3b5gqnflxx0auvu15nuVEteKvCd7upXa+p1KPw1d/SmWvb61dF79uTHQ8JBG35Gr7IkrIriUKq2kCqKQ6pMe19RlH6Nd115UYj58zBvENoAAFMSDIzMibdaZ68vO5G+gUG1psLcmVSX7Zkur0XvdNdIC197b79OdfbpYEuX2rr71d4zoL7kxM/sDSkvCo0KchMFv8pxzhH8UEgIbQCAnIqkVp2oqSia9r09/Um19wyovac/9Tryvi3tmPfe22/p6NNLJzuHz2Ur+FVOcK6iOKyKohDP9CHnCG0AgLxVHA6qOBycUeAbMlHwaxsOf+nnchP8hlr5SiNBlUVCKi0Kqqwo5L2PDL0PqrQopNJwkACIcRHaAABzWjaDX9uYcNc+heA3dK4/OfGI3bFKwqkgVxRUaWQk0JVFUvtFaUEvknbd8PVeMCwv8kJhaYQu4LmA0AYAwDlkGvycc+odGFR3X1KdfQPq7PVeu4ZeU8fSXzvG7Ld29+vYmW51DX/GwLSCYHE4MNLKN6qFL+3YmHOlkaHgN14wDLJSxywjtAEAkGNmNhz8qsoiWfvcvoFBL9j1JdXVO+a1b0AdvenBMKnO3tGv7T0DOt7WMxIY+5LqGzh3V/CQSCiQ1poXVEkkpJJwQKWRkEpSP29JxNsvDgdVEg4Ony+ODO0HVRIZeS2NBIevDQeNKWHSENoAAChQkVBAkVBEscmXyZ2W/uSguvpGWvk6ewfGtAqmhb+h46lruvsH1dOX1In2HnX3JdXTP6jufu+zJlu9YyLBgKk0HJww4I19LQ6nwmPq2uJIUKXjnUvbL6QVQAhtAABgWDgYULQkoGhJOKufOziY6iIeDnFJdfeN2U8dG7vf3T+g7r5k6tqkevqTOtHe7x1LHe/uT84oGIYCNhLwUqFuqKWvNDI6MG5eGdM7Lo5n9c9lWrX69s0AAGDeCATMaxWLBLUgi13E6QYHnXoGRoJcTyrknbXfn0y1BI7s96SOpe+/3NbvvU8dGxh0hDYAAIBMBQKm0og3cGIuKpyOXAAAgHmM0AYAAFAACG0AAAAFgNAGAABQAAhtAAAABYDQBgAAUAAIbQAAAAWA0AYAAFAACG0AAAAFgNAGAABQAAhtAAAABYDQBgAAUAAIbQAAAAWA0AYAAFAACG0AAAAFwJxzfteQU2bWLOngLHzVQkknZ+F7kBv8/gofv8PCx++w8PE7zNxK51zNeCfmfGibLWa2wzm3xe86MDP8/gofv8PCx++w8PE7zC26RwEAAAoAoQ0AAKAAENqy5xt+F4CM8PsrfPwOCx+/w8LH7zCHeKYNAACgANDSBgAAUAAIbQAAAAWA0JYhM7vazJ4zs/1mdrPf9WB6zGy5mf3czPaZ2R4z+7jfNWH6zCxoZjvN7Md+14LpM7OYmX3fzH6b+ndxm981YXrM7BOp/4Y+a2b3m1mx3zXNRYS2DJhZUNKdkt4kab2k95nZen+rwjQNSPqUc26dpK2SPsrvsCB9XNI+v4vAjH1Z0n845y6SVC9+lwXFzGolfUzSFufcRklBSe/1t6q5idCWmUsl7XfOHXDO9Ul6QNI1PteEaXDOHXPOPZN63y7vL4taf6vCdJhZXNJbJH3T71owfWZWKenVkv63JDnn+pxzZ/ytCjMQklRiZiFJpZKO+lzPnERoy0ytpMNp+03iL/yCZWarJF0s6Ql/K8E0fUnSn0ka9LsQzMgaSc2S7k51cX/TzMr8LgpT55w7IunvJB2SdExSq3PuZ/5WNTcR2jJj4xxjDpUCZGblkn4g6U+dc21+14OpMbPfl3TCOfe037VgxkKSNkv6mnPuYkmdkng+uICYWZW8XqbVkpZJKjOz6/ytam4itGWmSdLytP24aBIuOGYWlhfYvu2c+6Hf9WBaXiXpbWb2krzHE640s/v8LQnT1CSpyTk31ML9fXkhDoXjKkm/c841O+f6Jf1Q0mU+1zQnEdoy85Sk881stZlF5D14+aDPNWEazMzkPUuzzzl3h9/1YHqcc3/hnIs751bJ+/fvv51z/B9+AXHOvSzpsJldmDr0Okl7fSwJ03dI0lYzK039N/V1YjBJToT8LqCQOecGzOwmSQ/LGy1zl3Nuj89lYXpeJekDkhrNbFfq2Geccz/1sSZgvvkTSd9O/c/vAUkf8rkeTINz7gkz+76kZ+SNyN8plrPKCZaxAgAAKAB0jwIAABQAQhsAAEABILQBAAAUAEIbAABAASC0AQAAFABCG4B5zcySZrYrbcvabPxmtsrMns3W5wGY35inDcB81+2c2+R3EQBwLrS0Af+HvXuPk7Ou7/7//szszp5nk2w2yW4OEEiETDiGcFIKFSOC3oLe1YrK765UeX6DlQAAIABJREFUf9RaWx7lR3vTlgrFehe1tdVKH0otVm+t1ENRqiCKRa1KkCARTEJMCIcsOW025z3Pzuf3x3XN7uxmdjOTndlrZ/f1fDzmMdfhe13XZ4LIO9/r+l5fIA8ze9HMPmpmPw8/K8Ltp5jZD8zsmfB7Wbh9oZk9YGa/DD/ZaXziZvbPZrbJzL5nZnWR/SgAFY3QBmC2qxtze/QdOfuOuPtFkj4t6R/CbZ+W9EV3P0fSlyV9Ktz+KUk/cvdzFcydmZ0dZaWke9x9taRDkn6rzL8HwAzFjAgAZjUzO+bujXm2vyjpSnffYWbVkva4e4uZ7ZfU5u6D4fbd7j7fzDolLXH3/pxznCrp++6+Mlz/35Kq3f2vy//LAMw09LQBwPh8nOXx2uTTn7M8JJ4lBnCSCG0AML535Hw/Hi7/TNL14fK7Jf0kXP6BpN+XJDOLm1lyqooEMDvwNz4As12dmW3MWf+uu2df+1FjZk8o+AvuO8NtfyTpPjP7E0mdkm4Mt98s6V4ze6+CHrXfl7S77NUDmDV4pg0A8gifaVvr7vujrgUAJG6PAgAAVAR62gAAACoAPW0AAAAVgNAGAABQAQhtAAAAFYDQBgAAUAEIbQAAABWA0AYAAFABCG0AAAAVgNAGAABQAQhtAAAAFYDQBgAAUAEIbQAAABWA0AYAAFABCG0AAAAVgNAGAABQAQhtAAAAFYDQBgAAUAEIbQAAABWA0AYAAFABCG0AAAAVgNAGAABQAQhtAAAAFYDQBgAAUAEIbQAAABWA0AYAAFABCG0AAAAVgNAGAABQAQhtAAAAFYDQBgAAUAEIbQAAABWA0AYAAFABCG0AAAAVgNAGAABQAQhtAAAAFYDQBgAAUAEIbQAAABWA0AYAAFABCG0AAAAVgNAGAABQASIJbWZ2tZltNbPtZnZbnv3vN7NnzWyjmf3EzFLh9teb2VPhvqfM7Mqprx4AAGDqmbtP7QXN4pJ+Len1kjokPSnpne6+OadN0t2PhMvXSvqAu19tZudL2uvuu8zsLEmPuPviKf0BAAAAEYiip+0iSdvdfYe7D0i6X9J1uQ2ygS3UIMnD7U+7+65w+yZJtWZWMwU1AwAARKoqgmsulrQzZ71D0sVjG5nZH0i6RVJCUr7boL8l6Wl3789z7E2SbpKkhoaGC84888wSlA0AAFBeTz311H53b823L4rQZnm2HXeP1t3vkXSPmb1L0u2Sfmf4BGarJX1U0lX5LuDu90q6V5LWrl3rGzZsKEHZAAAA5WVmL423L4rbox2SluasL5G0a5y2UnD79C3ZFTNbIukBSf/L3Z8vS4UAAADTTBSh7UlJK81suZklJF0v6cHcBma2Mmf1TZK2hdvnSPqOpD9z959OUb0AAACRm/LQ5u5pSR+U9IikLZK+6u6bzOyucKSoJH3QzDaZ2UYFz7Vlb41+UNIKSX8Zvg5ko5ktmOrfAAAAMNWm/JUfU41n2gAAKN7g4KA6OjrU19cXdSkzUm1trZYsWaLq6upR283sKXdfm++YKAYiAACAaa6jo0NNTU069dRTZZZvDCFOlrurq6tLHR0dWr58ecHHMY0VAAA4Tl9fn1paWghsZWBmamlpKboXk9AGAADyIrCVz8n82RLaAAAAKgChDQAATDtdXV0677zzdN5552nRokVavHjx8PrAwEBB57jxxhu1devWgq/5uc99Tq2trcPXOe+884o6vtwYiAAAAKadlpYWbdy4UZJ05513qrGxUbfeeuuoNu4ud1cslr8P6vOf/3zR1333u9+tf/iHfxh3fzqdVlXVSHw6UQ25hoaGFI/Hi64pi542AABQMbZv366zzjpL73//+7VmzRrt3r1bN910k9auXavVq1frrrvuGm572WWXaePGjUqn05ozZ45uu+02nXvuubr00ku1b9++gq/56KOPat26dbr++ut1/vnn563hS1/6ks4++2ydddZZ+vM//3NJGr7u7bffrosuukg///nPJ/Xb6WkDAAAT+qv/3KTNu46U9Jyp9qTuePPqkzp28+bN+vznP6/PfOYzkqS7775b8+bNUzqd1mtf+1q97W1vUyqVGnXM4cOHdcUVV+juu+/WLbfcovvuu0+33Xbbcef+8pe/rB/+8IfD69mgtX79em3evFnLli3T9u3bR9XQ0dGh22+/XRs2bFBzc7PWrVunb3/727r66qt1+PBhrVmzRn/91399Ur81Fz1tAACgopx++um68MILh9e/8pWvaM2aNVqzZo22bNmizZs3H3dMXV2drrnmGknSBRdcoBdffDHvud/97ndr48aNw59EIiFJuvTSS7Vs2bK8NTzxxBO68sorNX/+fFVXV+td73qXfvzjH0uSEomE3vrWt5bkd9PTBgAAJnSyPWLl0tDQMLy8bds2ffKTn9TPf/5zzZkzRzfccEPe959lw5ckxeNxpdPpk77m2PWJZpeqq6sr2atT6GmbpJ0HevTuz63X4893RV0KAACzzpEjR9TU1KRkMqndu3frkUcemfIaLrnkEj322GPq6upSOp3W/fffryuuuKLk16GnbZKSddX66fYuXbaiVZee3hJ1OQAAzCpr1qxRKpXSWWedpdNOO02vec1rJnW+sc+0ffaznz3hMUuWLNFdd92l3/zN35S7681vfrPe9KY3Fd2bdyJMGF8Cl330v3T+srn6x3eeX9brAAAwVbZs2aJVq1ZFXcaMlu/PeKIJ47k9WgKptqQ27zocdRkAAGAGI7SVQKo9qR37u9UzUNpuUAAAgCxCWwmk2pJyl7buORp1KQAAYIYitJVAqj0pSdq8u7QvHgQAAMgitJXA4jl1StZWlfxt0QAAAFmEthIwM6Xak/S0AQCAsuE9bSWSamvWV37+soYyrnisNG8+BgBgturq6tLrXvc6SdKePXsUj8fV2toqKZgPNHeGg4ncd999euMb36hFixYdt++GG27QT3/6UzU3N0uSmpqa9N///d8l+gWlR2grkVR7Ur2DQ3phf7dWLGiMuhwAACpaS0uLNm7cKEm688471djYqFtvvbXo89x3331as2ZN3tAmSX//93+vt7zlLeMen06nVVVVNe56oceVAqGtRFJtI4MRCG0AAJTPF77wBd1zzz0aGBjQq1/9an36059WJpPRjTfeqI0bN8rdddNNN2nhwoXauHGj3vGOd6iurq7gHrrbb79dnZ2d2rFjhxYtWqQrrrhCjz76qI4dO6b+/n498sgjuvXWW/W9731PZqY77rhDb3vb2/Too4/q7rvv1vz587Vp0yY9++yzJf3dhLYSWbGgUdVx0+ZdR3Ttue1RlwMAQOk8fJu0p7QBRIvOlq65u+jDfvWrX+mBBx7Qz372M1VVVemmm27S/fffr9NPP1379+8fDkqHDh3SnDlz9I//+I/69Kc/rfPOOy/v+f74j/9Yd955pyTpnHPO0Re/+EVJ0tNPP60f//jHqq2t1ec+9zk9/vjj2rhxo+bOnat///d/1+bNm/XLX/5SnZ2duvDCC3X55ZdLktavX6/Nmzdr2bJlJ/GHMjFCW4kkqmJ61cImBiMAAFBGjz76qJ588kmtXRvM9NTb26ulS5fqDW94g7Zu3aqbb75Zb3zjG3XVVVcVdL7xbo9ed911qq2tHV6/6qqrNHfuXEnST37yE73rXe9SPB7XokWLdNlll2nDhg1KJBK69NJLyxLYJEJbSaXaknpsa2fUZQAAUFon0SNWLu6u3/3d39WHP/zh4/Y988wzevjhh/WpT31K3/jGN3Tvvfee9HUaGhrGXZ9o3vaxx5USr/wooVR7UvuP9Wvf0b6oSwEAYEZat26dvvrVr2r//v2SglGmL7/8sjo7O+Xuevvb366/+qu/0i9+8QtJwYjQo0dLO2PR5Zdfrvvvv19DQ0Pau3evfvrTnw73/JUTPW0lNDwYYdcRLTij9gStAQBAsc4++2zdcccdWrdunTKZjKqrq/WZz3xG8Xhc733ve+XuMjN99KMflSTdeOONet/73jfuQITcZ9ok6amnnjphDW9729u0fv16nXvuuTIzfeITn9CCBQtK+jvzsYm6+GaCtWvX+oYNG6bkWkf6BnXOnd/Tn159hj7wmyum5JoAAJTDli1btGrVqqjLmNHy/Rmb2VPunrfbjtujJZSsrdbSeXVMZwUAAEqO0FZiqTamswIAAKVHaCuxVFuzXtjfrZ6BdNSlAAAwKTP9EaooncyfLaGtxFLtSblLz+0p7UgVAACmUm1trbq6ughuZeDu6urqGvUeuEIwerTEUu0jI0jXLJsbcTUAAJycJUuWqKOjQ52dvH+0HGpra7VkyZKijiG0lVh7c62a66p5rg0AUNGqq6u1fPnyqMtADm6PlpiZBYMRGEEKAABKiNBWBqn2pJ7bc0RDGZ4DAAAApUFoK4NUW1J9gxm9sP9Y1KUAAIAZIpLQZmZXm9lWM9tuZrfl2f9+M3vWzDaa2U/MLJWz78/C47aa2RumtvLCZAcjbOIWKQAAKJEpD21mFpd0j6RrJKUkvTM3lIX+zd3PdvfzJH1M0ifCY1OSrpe0WtLVkv4pPN+0cnproxLxGIMRAABAyUTR03aRpO3uvsPdByTdL+m63Abunpt2GiRlHw67TtL97t7v7i9I2h6eb1pJVMW0cmEjgxEAAEDJRBHaFkvambPeEW4bxcz+wMyeV9DT9kdFHnuTmW0wsw1RvV8mO4KUlxICAIBSiCK0WZ5txyUbd7/H3U+X9L8l3V7ksfe6+1p3X9va2jqpYk/W6vakuroH1Hm0P5LrAwCAmSWK0NYhaWnO+hJJuyZof7+kt5zksZFJtTdLkjbxXBsAACiBKELbk5JWmtlyM0soGFjwYG4DM1uZs/omSdvC5QclXW9mNWa2XNJKST+fgpqLdmZbkyTxXBsAACiJKZ/Gyt3TZvZBSY9Iiku6z903mdldkja4+4OSPmhm6yQNSjoo6XfCYzeZ2VclbZaUlvQH7j401b+hEMnaai2bV88IUgAAUBKRzD3q7g9JemjMtg/lLN88wbEfkfSR8lVXOqm2pLbQ0wYAAEqAGRHKKNWe1Atd3eruT0ddCgAAqHCEtjJKtSXlLj2352jUpQAAgApHaCuj7HRWPNcGAAAmi9BWRm3NtZpTX80IUgAAMGmEtjIys2BmBHraAADAJBHayizVltRzu48oPZSJuhQAAFDBCG1llmpPqj+d0Ytd3VGXAgAAKhihrcyygxE28VwbAACYBEJbmZ3e2qhEPMZgBAAAMCmEtjKrjsf0qkWNDEYAAACTQmibAqm2pDbvOiJ3j7oUAABQoQhtUyDVllRX94D2He2PuhQAAFChCG1TINXeLEk81wYAAE4aoW0KrGprksR0VgAA4OQR2qZAU221Tmmpp6cNAACcNELbFGE6KwAAMBmEtimSakvqxa5uHetPR10KAACoQIS2KZJqT8pd2rqH3jYAAFA8QtsUyU5nxXNtAADgZBDapsiiZK3m1lfzXBsAADgphLYpYmZKtSfpaQMAACeF0DaFUm1JPbfnqNJDmahLAQAAFYbQNoVS7Un1pzN6YX931KUAAIAKQ2ibQqm2cDornmsDAABFIrRNodNaG5SoimkTz7UBAIAiEdqmUHU8pjMWNjEYAQAAFI3QNsWy01m5e9SlAACACkJom2Kp9qQOdA9o75H+qEsBAAAVhNA2xYZnRth9OOJKAABAJSG0TbEzFzVJYjorAABQHELbFGuqrdapLfW89gMAABSF0BYBprMCAADFIrRFINWW1ItdPTrWn466FAAAUCEIbRHIDkZ4jlukAACgQIS2CDCdFQAAKBahLQILkzWa15DguTYAAFCwSEKbmV1tZlvNbLuZ3ZZn/y1mttnMnjGzH5jZKTn7PmZmm8xsi5l9ysxsaqufPDMbnhkBAACgEFMe2swsLukeSddISkl6p5mlxjR7WtJadz9H0tclfSw89tWSXiPpHElnSbpQ0hVTVHpJpdqTem7PUaWHMlGXAgAAKkAUPW0XSdru7jvcfUDS/ZKuy23g7o+5e0+4ul7SkuwuSbWSEpJqJFVL2jslVZdYqi2pgXRGO/Z3R10KAACoAFGEtsWSduasd4TbxvNeSQ9Lkrs/LukxSbvDzyPuvqVMdZbV8HRWPNcGAAAKEEVoy/cMmudtaHaDpLWSPh6ur5C0SkHP22JJV5rZ5XmOu8nMNpjZhs7OzpIVXkqnzW9QoirGc20AAKAgUYS2DklLc9aXSNo1tpGZrZP0F5Kudff+cPNbJa1392PufkxBD9wlY49193vdfa27r21tbS35DyiFqnhMZy5q0qZdTBwPAABOLIrQ9qSklWa23MwSkq6X9GBuAzM7X9JnFQS2fTm7XpZ0hZlVmVm1gkEIFXl7VAqea9u864jc83Y0AgAADJvy0ObuaUkflPSIgsD1VXffZGZ3mdm1YbOPS2qU9DUz22hm2VD3dUnPS3pW0i8l/dLd/3Nqf0HppNqTOtgzqD1H+qIuBQAATHNVUVzU3R+S9NCYbR/KWV43znFDkn6vvNVNnVTbyGCEtua6iKsBAADTGTMiROjMNkaQAgCAwhDaItRYU6VTW+oZQQoAAE6I0BaxVDvTWQEAgBMjtEVsdXuzXurq0dG+wahLAQAA0xihLWLZwQjP7TkacSUAAGA6I7RFjOmsAABAIQhtEVvQVKOWhgShDQAATIjQFjEzYzACAAA4IULbNJBqS2rr3qMaHMpEXQoAAJimCG3TQKo9qYF0Rjs6u6MuBQAATFOEtmlgeDqr3YcjrgQAAExXhLZpYPn8BtVUxRiMAAAAxkVomwaq4jGduahJmwhtAABgHIS2aSI7gtTdoy4FAABMQ4S2aSLVltShnkHtPtwXdSkAAGAaIrRNE8yMAAAAJkJomybOWJSUmXjJLgAAyIvQNk001lTp1JYGetoAAEBehLZpJNXGdFYAACA/Qts0kmpP6uUDPTrSNxh1KQAAYJohtE0j2ZkRntt9NOJKAADAdFNVSCMzm1dAs4y7H5pkPbPa6uERpId10fJC/sgBAMBsUVBok7Qr/NgEbeKSlk26olmstalG8xsTPNcGAACOU2ho2+Lu50/UwMyeLkE9s5qZaRWDEQAAQB6FPtN2aYna4ARS7Un9es8xDQ5loi4FAABMIwWFNnc/4dxKhbTBiaXakhoYyuj5zmNRlwIAAKaRE4Y2M3u9mf2zmZ0Xrt9U/rJmr9VMZwUAAPIopKftA5L+RNINZnalpPPKW9Lstnx+o2qrY4Q2AAAwSiGhrdPdD7n7rZKuknRhmWua1eIx0xmLGIwAAABGKyS0fSe74O63Sfpi+cqBFDzXtmnXEbl71KUAAIBp4oShzd2/lV02s22SXmtmd5jZdWZ2ahlrm7VS7Ukd7h3UrsOM7QAAAIFip7H6rKQ9krokXSPpV2b2rJndZWbVJa9ulspOZ8VzbQAAIKvY0HaDu3/A3T/t7u+XdJmk/5J0RNInSl7dLHXmoiaZEdoAAMCIYkPbYTM7J7vi7hslXeLufyvpNSWtbBZrqKnS8pYGbd59OOpSAADANFHoNFZZvyfpy2a2UdJGSWdIyr66P1HKwma7Ve1JPdNxKOoyAADANFFUT5u7PyfpIknflbRA0nZJ/8PMGiTdX/ryZq9UW1I7D/TqcO9g1KUAAIBpoNieNrn7kKSvhZ9cf12SiiApGEEqSc/tPqKLT2uJuBoAABC1onrazOxKM/sXM/s7M7vRzC4ws5piL2pmV5vZVjPbbma35dl/i5ltNrNnzOwHZnZKzr5lZvY9M9sStjm12OtXgtXZEaS8ZBcAAKj4gQhfkvRtSeslnSbpQ5I2FXMCM4tLukfBK0NSkt5pZqkxzZ6WtNbdz5H0dUkfy9n3RUkfd/dVCm7V7ivyN1SE1qYazW9MMIIUAABIKv726HZ3fyBcHnt7tFAXhefZIUlmdr+k6yRtzjZw98dy2q+XdEPYNiWpyt2/H7Y7dpI1THtmplR7Mz1tAABAUvE9bT8ysz82M5vENRdL2pmz3hFuG897JT0cLr9K0iEz+w8ze9rMPh723I1iZjeZ2QYz29DZ2TmJUqOVaktq295jGkhnTtwYAADMaMWGttWSfl/SbjP7jpl9xMzeXuQ58gW+vJNsmtkNktZK+ni4qUrSb0i6VcHE9adJes9xJ3O/193Xuvva1tbWIsubPlLtSQ0MZfR854ztUAQAAAUqKLSZ2R3h4t9JukDSckl3SNom6eIir9khaWnO+hJJu/Jcc52kv5B0rbv35xz7tLvvcPe0pG9KWlPk9SsG01kBAICsQnvavhd+36zgGbNfSrpdQU/X+iKv+aSklWa23MwSkq6X9GBuAzM7X8E8p9e6+74xx841s2z32ZXKeRZuplk+v0G11TGeawMAAIUNRHD3x8PFP3X3F8PXfKyWdLaCnravF3pBd0+b2QclPSIpLuk+d99kZndJ2uDuDyq4Hdoo6Wvh43Mvu/u17j5kZrdK+kH4XN1Tkv650GtXmnjMdOaiJD1tAACg6NGj/yFpTXi78heSfmFmlxR7UXd/SNJDY7Z9KGd53QTHfl/SOePtn2lS7Ul955ndcndNbvwHAACoZIU+0/bbZna3pCYzWzVmxOa95SkNUvBc2+HeQb1yqDfqUgAAQIQKfabtpwqeHZsr6ROStpnZL8zs25JIE2WUnc6KW6QAAMxuhT7T9oqkL5rZ8+7+U0kys3kKRpE+V8b6Zr0zFzXJLJjO6qrVi6IuBwAARKSoZ9qygS1cPiDpQMkrwij1iSotn99ATxsAALNcUaHNzFok/bakPgVzjj7r7tweLbNUW1Ibdx6KugwAABChYmdEeEBSq6T/o+C1HIfNjNujZZZqT6rjYK8O9w5GXQoAAIhIsaGtyd3vkrTX3a+Q9E5Jny99WciVnRlhCy/ZBQBg1io2tPWF3/1mVufu35D0xhLXhDEYQQoAAIp9ue7fhqNG/13SfWb2M0mLS18Wci1oqtX8xhqmswIAYBYruKfNzGKSznD3A+7+CQUzGiyVdF25isOIVDvTWQEAMJsVHNrcPSPpdTnr/9fd/9TdN5WlMoySaktq276jGkhnoi4FAABEoNhn2jaa2R1hrxum0Or2pAaHXNv3HYu6FAAAEIFiw9dSSddL2mVm3zKzD5vZ28tQF8YYHozAc20AAMxKxc6I8NuSZGY1klZLOlvSRZK+VvrSkOvUlgbVVceD59ouiLoaAAAw1YodPSpJcvd+Sb8IP5gC8ZjpzLYmbd59OOpSAABABAq6PWpmJwxnhbTB5KTaghGk7h51KQAAYIoV2tO2ysyemWC/SWouQT2YQKo9qS8/8bJeOdSrJXProy4HAABMoUJD25kFtBmaTCE4sex0Vpt2HSG0AQAwyxQU2tz9pXIXghM7c1FSMQums3rD6kVRlwMAAKYQ71urIHWJuJbPb+C1HwAAzELFTGNlZra0nMXgxFLtzUxnBQDALFTMNFYu6ZtlrAUFSLUl9cqhXh3uGYy6FAAAMIWKvT263swuLEslKAgzIwAAMDsVG9peK+lxM3vezJ4xs2dP8CoQlFh2BCmhDQCA2aXYGRGuKUsVKFhrU41am2p4rg0AgFmmqJ628NUfcyS9OfzM4XUgUy/VlqSnDQCAWaao0GZmN0v6sqQF4edLZvaH5SgM40u1J7V931ENpDNRlwIAAKZIsbdH3yvpYnfvliQz+6ikxyX9Y6kLw/hSbUkNDrm27Tuq1e3MHgYAwGxQ7EAE0+jpqobCbZhCq7MjSHmuDQCAWaPYnrbPS3rCzB4I198i6V9KWxJO5JSWBtUn4jzXBgDALFJwaDMzk/Q1ST+UdJmCHrYb3f3p8pSG8cRjpjMXNdHTBgDALFJwaHN3N7NvuvsFkn5RxppQgFR7Ut/auEvuriBPAwCAmYwZESpUqq1ZR/vS6jjYG3UpAABgCjAjQoViOisAAGaXYp9pe78kXqY7DZyxsEkxkzbtOqI3rF4UdTkAAKDMin2m7e/DZ9oQsbpEXKe1NjIYAQCAWYJn2ipYqi2pLdweBQBgVjiZZ9rWT/aZNjO72sy2mtl2M7stz/5bzGxzeI0fmNkpY/YnzewVM/t0sdeeSVLtSb1yqFeHegaiLgUAAJRZsS/XvWayFzSzuKR7JL1eUoekJ83sQXffnNPsaUlr3b3HzH5f0sckvSNn/4cl/WiytVS6VNvIYIRXnz4/4moAAEA5FdTTZmZ/Kknu/pKki9z9pexH0u8Vec2LJG139x3uPiDpfknX5TZw98fcvSdcXS9pSU4tF0haKOl7RV53xlnVxnRWAADMFoXeHr0+Z/nPxuy7ushrLpa0M2e9I9w2nvdKeliSzCwm6e8k/clEFzCzm8xsg5lt6OzsLLK8ytHaVKMFTTW89gMAgFmg0NBm4yznWy/mXFmet6HZDZLWSvp4uOkDkh5y95352g+fzP1ed1/r7mtbW1uLLK+ypNqT9LQBADALFPpMm4+znG/9RDokLc1ZXyJp19hGZrZO0l9IusLd+8PNl0r6DTP7gKRGSQkzO+buxw1mmC1SbUn9ZNt+9aeHVFMVj7ocAABQJoWGtnPN7IiCXrK6cFnhem2R13xS0kozWy7pFQW3Xt+V28DMzpf0WUlXu/u+7HZ3f3dOm/coGKwwawObFPS0pTOubXuP6azFzVGXAwAAyqSg0ObuJevCcfe0mX1Q0iOS4pLuc/dNZnaXpA3u/qCC26GNkr4WTob+srtfW6oaZpLcEaSENgAAZq5iX/lREu7+kKSHxmz7UM7yugLO8a+S/rXUtVWaU1saVJ+I81wbAAAzXLEv18U0E4uZVrUlGUEKAMAMR2ibAVJtSW3ZdUTuxY4JAQAAlYLQNgOk2pM62p9Wx8HeqEsBAABlUlRos8ANZvahcH2ZmV1UntJQqOxghE081wYAwIxVbE/bPyl4V9o7w/WjCuYRRYTOWNSkmInn2gAAmMGKHT16sbuvMbM7RyYkAAAgAElEQVSnJcndD5pZogx1oQi11XGd3tqozbsOR10KAAAok2J72gbNLK5wFgQza5WUKXlVKBrTWQEAMLMVG9o+JekBSQvM7COSfiLp/5S8KhQt1ZbUrsN9Otg9EHUpAACgDAq+PWrB1AQ/lvSUpNcpmMLqLe6+pUy1oQip9mAwwpbdR/TqFfMjrgYAAJRawaHN3d3MvunuF0h6row14SSsypnOitAGAMDMU+zt0fVmdmFZKsGkzG+s0cJkDc+1AQAwQxU7evS1kn7PzF6S1K3gFqm7+zklrwxFSzGdFQAAM1axoe2aslSBkki1J/Xf2/arb3BItdXxqMsBAAAlVNTtUXd/SdIRSQslnZLzwTSQamtWOuPavu9Y1KUAAIASK6qnzczeJ+lmSUskbZR0iaTHJV1Z+tJQrOwI0s27juisxc0RVwMAAEqp2IEIN0u6UNJL7v5aSedL6ix5VTgpp8yrV30iznNtAADMQMWGtj5375MkM6tx9+cknVH6snAyYjHTqjZmRgAAYCYqNrR1mNkcSd+U9H0z+5akXaUvCycrO4I0k/GoSwEAACVU7ECEt7r7IXe/U9JfSvoXSdeVozCcnNXtSR3rT6vjYG/UpQAAgBIqdiDCh/JsPk/SXaUpB5M1PBhh92Eta6mPuBoAAFAqxd4e7c75DCl4b9upJa4Jk/CqhU2Kx4zn2gAAmGGK6mlz97/LXTezv5X0YEkrwqTUVsd1emsDI0gBAJhhiu1pG6te0mmlKASlk2pLahM9bQAAzCjFPtP2rKTssMS4pFbxPNu0k2pP6psbd+lA94DmNSSiLgcAAJRAsXOP/o+c5bSkve6eLmE9KIFUWzAbwpbdR/SaFfMjrgYAAJRC0XOP5nxeIbBNT6vamiSJwQgAAMwgxd4evWWi/e7+icmVg1JoaazRomQtgxEAAJhBir09ulbB3KPZEaNvlvRjSTtLWRQmL9XOdFYAAMwkxYa2+ZLWuPtRSTKzOyV9zd3fV+rCMDmptqR+9OtO9Q0OqbY6HnU5AABgkop95ccySQM56wPi5brTUqo9qaGMa9veY1GXAgAASqDYnrb/K+nnZvZAuP5WSV8sbUkohVTbyHRWZy9pjrgaAAAwWcXOiPARM3tY0m8oeF/be9z96bJUhklZNq9eDYk4z7UBADBDFHR71MwuNLNFkuTuv5B0SNI6STea2bwy1oeTFIuZVrUlGUEKAMAMUegzbZ9V+CybmV0u6W8kfUHSYUn3lqc0TFaqPaktu48qk/ETNwYAANNaoaEt7u4HwuV3SLrX3b/h7n8paUV5SsNkpdqSOtaf1s6DPVGXAgAAJqng0GZm2effXifpv3L2FTuYAVMk1R4ORuC5NgAAKl6hoe0rkn5kZt+S1CvpvyXJzFYouEWKaehVC5sUjxnPtQEAMAMUFNrc/SOS/j9J/yrpMnfPPiQVk/SHxV7UzK42s61mtt3Mbsuz/xYz22xmz5jZD8zslHD7eWb2uJltCve9o9hrzya11XGtaG2kpw0AgBmg4Fub7r4+z7ZfF3tBM4tLukfS6yV1SHrSzB509805zZ6WtNbde8zs9yV9TMGzdD2S/pe7bzOzdklPmdkj7n6o2Dpmi1R7Uut3dEVdBgAAmKRiZ0QohYskbXf3He4+IOl+SdflNnD3x9w9+/T8eklLwu2/dvdt4fIuSfsktU5Z5RUo1ZbU7sN9OtA9cOLGAABg2ooitC3W6AnmO8Jt43mvpIfHbjSziyQlJD2fZ99NZrbBzDZ0dnZOstzKxmAEAABmhihCm+XZlvdFYmZ2g6S1kj4+Znubgim1bnT3zHEnc7/X3de6+9rW1tndEbcqZzorAABQuYp6XYeZ1Uj6LQWTxA8f6+53FXGaDklLc9aXSNqV51rrJP2FpCvcvT9ne1LSdyTdnu85O4w2ryGhtuZaetoAAKhwxb5j7VsKXvHxlKT+E7Qdz5OSVprZckmvSLpe0rtyG5jZ+QpmYbja3fflbE9IekDSF939ayd5/VknxXRWAABUvGJD2xJ3v3oyF3T3tJl9UNIjkuKS7nP3TWZ2l6QN7v6ggtuhjZK+ZmaS9LK7XyvptyVdLqnFzN4TnvI97r5xMjXNdKn2pH746071DQ6ptjoedTkAAOAkFBvafmZmZ7v7s5O5qLs/JOmhMds+lLO8bpzjviTpS5O59myUaktqKOP69d6jOmfJnKjLAQAAJ6HYgQiXKXg32tbw5bbPmtkz5SgMpcMIUgAAKl+xPW3XlKUKlNXSufVqrKniuTYAACpYUaHN3V8ys7mSVkqqzdn1UkmrQknFYqZVbU30tAEAUMGKfeXH+yTdrOA1HRslXSLpcUlXlr40lFKqLamvP9WhTMYVi+V7VR4AAJjOin2m7WZJF0p6yd1fK+l8SbN7yoEKkWpPqntgSC8f6DlxYwAAMO0UG9r63L1PCl606+7PSTqj9GWh1FJtzZLEc20AAFSoYkNbh5nNkfRNSd83s28pz2wGmH5WLmxUPGY81wYAQIUqdiDCW8PFO83sMUnNkr5b8qpQcrXVca1c0EhPGwAAFaqonjYL3GBmH3L3HykYjHBeeUpDqaXakvS0AQBQoYq9PfpPki6V9M5w/aike0paEcom1Z7UniN96jp2stPGAgCAqBQb2i529z+Q1CdJ7n5QUqLkVaEsUm3hzAjcIgUAoOIUG9oGzSwuySXJzFolZUpeFcpiVRvTWQEAUKmKDW2fkvSApIVm9hFJP5H0NyWvCmUxtyGh9uZaetoAAKhAxY4e/bKZPSXpdeGm68J3taFCpNoZjAAAQCUqKLSZ2YNjN4XfbzAzufu1pS0L5ZJqS+q/ntunvsEh1VbHoy4HAAAUqNCetksl7ZT0FUlPaCS0ocKk2pPKuLR1z1Gdu3RO1OUAAIACFfpM2yJJfy7pLEmflPR6Sfvd/Ufh+9pQIZjOCgCAylRQaHP3IXf/rrv/jqRLJG2X9EMz+8OyVoeSWzK3Tk01VTzXBgBAhSl4IIKZ1Uh6k4IX656qYCTpf5SnLJRLLGZa1Zakpw0AgApT6ECELyi4NfqwpL9y91+VtSqUVao9qa9u2KlMxhWL8XgiAACVoNCetv9HUrekV0n6I7Ph/9CbJHf3ZBlqQ5mk2pLqGRjSSwd6tHx+Q9TlAACAAhQU2ty92JfwYhpLtY/MjEBoAwCgMhDGZqEVCxpVFTNt3n046lIAAECBCG2zUG11XCsWNDKCFACACkJom6VSjCAFAKCiENpmqVR7UnuP9Gv/sf6oSwEAAAUgtM1SuYMRAADA9Edom6VSbWFo4xYpAAAVgdA2S82pT2jxnDp62gAAqBCEtlmM6awAAKgchLZZLNWe1I7OY+odGIq6FAAAcAKEtlks1ZZUxqWte49GXQoAADgBQtsstpoRpAAAVAxC2yy2ZG6dmmqqmM4KAIAKQGibxcxMq9qT9LQBAFABCG2zXKotqef2HNVQxqMuBQAATCCS0GZmV5vZVjPbbma35dl/i5ltNrNnzOwHZnZKzr7fMbNt4ed3prbymSfVnlTPwJBe6uqOuhQAADCBKQ9tZhaXdI+kaySlJL3TzFJjmj0taa27nyPp65I+Fh47T9Idki6WdJGkO8xs7lTVPhMxMwIAAJUhip62iyRtd/cd7j4g6X5J1+U2cPfH3L0nXF0vaUm4/AZJ33f3A+5+UNL3JV09RXXPSCsXNqoqZjzXBgDANBdFaFssaWfOeke4bTzvlfTwSR6LE6ipimvFgkZ62gAAmOaiCG2WZ1vep+DN7AZJayV9vJhjzewmM9tgZhs6OztPutDZIsUIUgAApr0oQluHpKU560sk7RrbyMzWSfoLSde6e38xx7r7ve6+1t3Xtra2lqzwmSrVltS+o/3qPNp/4sYAACASUYS2JyWtNLPlZpaQdL2kB3MbmNn5kj6rILDty9n1iKSrzGxuOADhqnAbJiEVzoywhVukAABMW1Me2tw9LemDCsLWFklfdfdNZnaXmV0bNvu4pEZJXzOzjWb2YHjsAUkfVhD8npR0V7gNk7C6rVmStIlbpAAATFtVUVzU3R+S9NCYbR/KWV43wbH3SbqvfNXNPs311Vo8p47BCAAATGPMiABJ2cEIzEEKAMB0RWiDpGAwwo793eoZSEddCgAAyIPQBklBT5u7tHXP0ahLAQAAeRDaIInprAAAmO4IbZAkLZlbp6baKl6yCwDANEVogyTJzJRqS9LTBgDANEVow7BUe1LP7T6qoUzeWcUAAECECG0YlmpLqndwSC92dUddCgAAGIPQhmHZ6ax4rg0AgOmH0IZhKxc0qTpuPNcGAMA0RGjDsERVTCsWNNHTBgDANERowyiMIAUAYHoitGGUVHtSnUf7te9oX9SlAACAHIQ2jJKdGWHLbqazAgBgOiG0YZTh6ax4rg0AgGmF0IZRmuurtXhOnTbtOhx1KQAAIAehDcdZ3c5gBAAAphtC22QNDUpPfUF65SlpsDfqakoi1Z7UC/u71TOQjroUAAAQqoq6gIq3f5v0n38ULFtMmv8qadHZwWfhWdKic6TG1mhrLFKqLSl36bk9R7Vm2dyoywEAACK0TV7rmdIfbZT2PDvyeXm99OzXRto0LhoJcovCIDfvNCkWj67uCeROZ0VoAwBgeiC0TVYsJs1bHnxS145s7zkg7f1VGOTC7x2PSZnwlmN1vbRw9egeuYUpKdEQze/IsXhOnZK1VTzXBgDANEJoK5f6edLyy4NPVrpf6twaBLhsoPvVN6QN94UNTGpZEfbGnR0EuUVnS40LJbMpK93MlGpP8toPAACmEULbVKqqkdrOCT5Z7tLhnTk9cs9Ir/xC2vTASJuG1rA3LifItayQ4uX7x5dqa9a//fwlDWVc8djUBUYAAJAfoS1qZtKcZcHnzDeNbO89JO3dFPbIPROEuic+Iw0NBPuraqUFq0YHuYWrpZqmkpSVak+qbzCjF/Z3a8WCxpKcEwAAnDxC23RVN0c69TXBJ2toMBituufZkSC35dvSL7440mbu8tFBbtFZUnJx0bdXh2dG2H2E0AYAwDRAaKsk8epgsMLClHTuO4Jt7tLR3aOD3J5npS0PjhxXN3dMj9xZUusZwfnGsWJBo6rjps27jujac9vL/MMAAMCJENoqnZmUbA8+r3rDyPb+o9LezaOD3JOfk9J9wf54InhdSW6P3MKzgh4+SYmqmFYuaGIEKQAA0wShbaaqaZKWXRx8sobS0oHnR/fKbXtE2vilkTZzlg0Hubc3Nuo/Xpkn+YVTOnoVAAAcz9w96hrKau3atb5hw4aoy5jeju4dCXLZV5Hs3yYp+N9GpiapWMvpUsOCYHaHhgVS44JgVGvjwpHlurmEOwAAJsHMnnL3tfn20dMGqWlh8Fm5bmTbQLd+9fR6/duD39EfrOjRYt8nHd0l7f6l1N0p+dDx54lVh0EuG+wW5gl5C4L1urnBi4kBAEBBCG3IL9Ggpedcrn97oE9LlpyhD/zmipF9mYzUe1Dq3icd2xeEuGN7c5b3Bfv2bgq+M3kmno9VBSEuN8iN14tXN4+ABwCY9QhtGFdzXbWWzK07fmaEWExqaAk+C1ZNfBL3MOCNE+yOdQbf+54L9mcGjz+HxaWG+WOCXRjqxoa9+pZpO6crAACTQWjDhFJtk5zOyiyY0qt+XvCakYm4S32HRoLccQEv/OzfFnwP9ee5XiwIbtkQN/a27KiAN7+ss0oAAFBK/BcLE1rd3qzvb9mr7v60GmrK/D8Xs+BZt7q5UuurJm7rLvUfyRPwxvTmHXg+aJPuzXfBIOBlb8HmDqI4bkDFePsmeUzZz5e7PRZ8YvGR5eGPBT2aY7cPt7Ux+8a2tXHOO87nuLY2znljQc9u3vPEjz9++LzxnHXLsy27PrbusW0YWANg+iC0YUKp9qTcpef2HNUFp8yNupwRZlJtc/CZv2Litu7SwLGRnrrjevE6g1u4mUz2gNHHjj7ZOPvGtBtvX8nPN+Z0E53PPRhA4pkxH5cy+bYPhceM2Z5te/zFZ6BxwuCoMDk26OULp/H8x+Q7bqJjqmql6lqpqm7Md61UXVf4dzxBIC039+B53nR/MP3g0GDw701VTfip5VEOFI3Qhgml2kems5pWoa0YZsF762qapJbTo65m5nDPCXX5Ql8mCMJ5w2BOYBwbBgv5DLf148+b7zzD23Lb+DjHDI0JsmPb5Dsuu57nN46tZ+wxw9canPg3DA0GASDdKw32Bd+eOfE/p7xs/FA3HAyLDIIn+i5HQMlkwkDUP/LnMzQw8kkPnGB/uP24/dltuec40TFjrzmgE/7FJhvEh4NcGObiiTHb822rkeI5+wtqO861ZmuAH/53byj4zqTD5UzOcnZ7+O9lTWPwMvuIENowofbmWjXXVeu/tuzVZSvm69SWetls/Rcco5mF/2cfE/9XEgH3MDTkhLhR3+FnsPcE33mO7enKvz3fc6SFilXnhLjanICYE+zciwtg+V49NCk2EnDi1eFydbiekKoS4XdN8JfAUfuz7RMjn2z77H6z4Hek+8Lf1BcG8f5xtvUHdwGy+9M5+4f6R2a4maz4iULj2ICYGPnnN9w2/J3Zv5Bk0uFfYrLLueGoyKA0anvOsQVfJ7dNzvlO5i8951wv/c/PlubP/SRE8v+0Zna1pE9Kikv6nLvfPWb/5ZL+QdI5kq5396/n7PuYpDcp+C/F9yXd7DP9DcERMjNd8apWPfjLXXps6w+1MFmjS05r0SWnteji5fO0fH4DIQ6Igln4H89E8JjAVMhkigiDE32P6TEc7JV6DwS3gbMh57hQlBuCcsJTKfdX2sAk95Hev+Eg1z9BGBwY+eeXLyCOd/xgb054zNM232udRgmfHY1VBb2LsfBjud9V4aMA2eV4nrbZ7dUTn294e2x0m3HPFxtzzdxaxlxnzilT8o92PFP+v1Azi0u6R9LrJXVIetLMHnT3zTnNXpb0Hkm3jjn21ZJeoyDMSdJPJF0h6YflrXp2++T15+nmdSu1fkeX1u84oJ8936VvbdwlSVrQlBPiTpun0whxwMwVi0mJ+uCD6JmN9IRFKTM0EuQsX1DiPZulEsVfKy6StN3dd0iSmd0v6TpJw6HN3V8M943tu3RJtZISkkxStaS95S95djMznd7aqNNbG/Xui0+Ru+uF/d1av+NAGOS69OAvgxDXOhzi5uni5S06vZUQBwAzWiweBnnCfLlFEdoWS9qZs94h6eJx2o7i7o+b2WOSdisIbZ929y1j25nZTZJukqRly5ZNumCMZmY6rbVRp7U26l0XL5O768WunuEAt35Hl/4zDHHzG2t0yWnzhoPc6a2NhDgAAE5CFKEt33+xC3omzcxWSFolaUm46ftmdrm7/3jUydzvlXSvFEwYP4laUQAz0/L5DVo+v0HvvCgIcS+NCnEH9O1ndkuS5jcmdHF4O/WS5fO0YgEhDgCAQkQR2jokLc1ZXyJpV4HHvlXSenc/Jklm9rCkSyT9eMKjMKXMTKfOb9Cp8xt0fRjiXj7QMxzg1u/o0ndyQ9zy8HbqaS1aSYgDACCvKELbk5JWmtlySa9Iul7Suwo89mVJ/6+Z/Y2CHrsrFIwyxTRmZjqlpUGntDToHRcGIW7ngd5Rt1O/82wQ4loaEro4vJ168fIgxMVihDgAACyKt2WY2RsVhK24pPvc/SNmdpekDe7+oJldKOkBSXMl9Una4+6rw5Gn/yTpcgW3VL/r7rdMdK21a9f6hg0byvlzMEnuro6DvXo8DHBP7DigVw4F007Na0jo4uXzhkenvmpBEyEOADBjmdlT7r42776Z/oozQltl2jnmdmo2xM2trx51O/WMhYQ4AMDMMVFoq7A3CWK2WDqvXkvn1evta4PHH3ce6NETL4y8YuS7m/ZIkubUVw/3xF1CiAMAzGCENlSEbIh72wXBwOGOgz16IvueuBe69Mim4HV9c+qrddGpIyHuzEWEOADAzEBoQ0VaMrdeSy6o12+FIe6VQ716IucVI9/bHIS45rpqXbR85D1xqxYlCXEAgIrEM22YkXYd6tUTL3Rp/fMHtP6FLr3U1SNJStZW6aLwmbhLTmvRqrak4oQ4AMA0wUAEQtust/tw78jt1B1dejEMcU21Vbp4+TxdcMo8LZ9fryVzg9uwzXXVEVcMAJiNCG2ENoyx53Bf0BMX3k59YX/3qP3NddVaOq9OS8MQt3RevZbOrdPSefVaPKdOtdXxiCoHAMxkhDZCG07gcM+gdh7s0c4DPeF3r3Ye7NHLB3rUcbBXA+nMqPYLkzVaNq9eS+fWa0lOoFs2r14Lk7XccgUAnBRe+QGcQHN9tZrrm3XW4ubj9mUyrs5j/dp5IAhx2UCXfZfc7o2vKPfvPtVx0+I5QYhbMjcIcrm9dnPrq5mqCwBQNEIbcAKxmGlhslYLk7Vae+q84/YPpDPadah3uIfu5bC3ruNAjx7ZtUcHugdGtW+sqdKSsGcuCHJ1YbCr15K5dapP8K8lAOB4/NcBmKREVUynzm/QqfMb8u4/1p8Obrse6NHOg73Dyy91desn2/ard3BoVPv5jYm8PXTL5tWrrblWVfHYVPwsAMA0Q2gDyqyxpkqr2pJa1ZY8bp+7q6t7YPjWa0c21B3s0dM7D+o7z+7WUGbk3ms8Zmprrs3TQxestzbWcOsVAGYoQhsQITPT/MYazW+s0fnL5h63Pz2U0e7DfSODJHKep3tsa6c6j/aPal9bHRsZ8Tq3blSgW9BUq7n11fTUAUCFIrQB01hVPDb8yhGdfvz+3oEhdRzMGfGaHSxxsFdPvnBAR/vTo9qbBa8zmdeQUEtDQvMaEprXUDO83NKY3ZZQS0ON5jUklKgi5AHAdEBoAypYXSKulQubtHJh03H73F2HeweHb7vuP9avrmMDOtAdfLq6+/XC/m499dJBHegeUGact/801VRpXmNi3KA3r3Fke0tDjeoSvMMOAMqB0AbMUGamOfUJzalP6JwlcyZsm8kEAa+rOxvq+oPlYwM52wb0yqE+PfvKYR3oHtDgUP6UV1cdz9NrN37Qa6yp4jk8ACgAoQ2AYjHT3IaE5jYkCmrv7jranx4T6kaCXtCTN6CuYwPatveYurr71TeYyXuuRDw2Eu7yBL3c7S0NCSVrqxXj5cUAZiFCG4CimZmStdVK1laP+6qTsXoG0mNuz+YPei919ehA94COjXkeLyseM82tTxzXazenPqGmmio11VapqbZaTbVVaqytUjJcb6ypUn0iTq8egIpFaAMwJeoTVaqfVxUMqihA3+CQDvYM5A16B7pHtm/ZdURd3QM63Dt4wnPGY6bGMNg11lQpmRPucsNeEP6qh9tll5O11WqoiTMCF0AkCG0ApqXa6rjamuvU1lxXUPtMxnVsIK1jfWkd7UvraN9g8N0/snwsz/Y9R/p0dF9ax8L18Z7Vy1WfiOcNdI05PX3ZIJjb05cbDGuqYvT6ASgKoQ3AjBCLjdyyPVnurv50ZnTo60vrWP+gjmSXc/Yd60/rSLi861BvGPzS6hkYOuG1quM2pjcvDHR5gl92e30irvpEleoS8XA5WOe1LMDsQGgDgJCZqbY6rtrquFqbak76POmhjLr7h4YD3bH+3BA4GPbyBcsjPYPBdGfHcvaN9xqWsapilhPkqlRXHVdDTVx1iSrVVwfb6xJxNdQE+7KBry5RNSr8ZdvVJ+Kqr65SfU1c1dwKBqYNQhsAlFhVPKbm+pia6yfX69c7ODSq1693YEg9A0PqHhhZ7h0cUs9AWt39Q8G2wSH1DgS9fYd7B7XncG/QLqd9MarjFga9MODVBIGuLhEfFfIaErnbgrA4HBwT8ZywWBWeg2cDgWIR2gBgGjKzMChVaWGytmTnzWQ8DHrZkJcT+AbSw/u6+9M5ITAMhgMjy4d6BrTrUNC2JwyJ/en8r3UZTyIeGxX+6sJeztrqmGqqgu/aqrhqxm6rjqu2KviuCdsML1fHVRPuG9WuKkZIRMUjtAHALBKLmRpqqtRQU/r/+x8aDoRp9fRne/bSYbAbCXe9w+ujt/UNDqkvPaS+wYwO9QyqbzAIgn2DGfWH+woZKDKeqpiNCoUjgW8k5I0EvmxIHL2t4FAZBkaCIkqJ0AYAKInsK1Uaa6qk42dWK4mhjKs/DHZ9g2HQG8yMbEsPBQEvd9tg7r5MGAzD5eFgGDyD2DdmW/9gRgNDxfUg5orHbLi3r7Y6rkRVTDVVsTHfcSXiMdVUx3K+42PWY6oJA2TN2GMnXA+OZaTyzEBoAwBUjHgse9t46q6ZDYojgW8kMGbD3UhIHL1t7PrAUNBrGHwHvYiHewc1kM6oP50Z8z25nsVciaqYasYEwGwgPKnwOE676niwLVE18l0dt/A72MaMJieP0AYAwASiCIpZmYwHAS8MceOFu+x6vjbjHzs0vN43mNGR3vQ4x5cuPErBberhEDdBuMvuqx7en902+vjqeBA8J2x3wrbBd3yaB0pCGwAA01QsZqqNBbdWpZMfjTxZhYbHwaFgeWDINZDODK8Phsfmro/XbiBcPtafHrPPjztHutD34hQoZho/3MVjWrdqgW656oySXrMYhDYA+P/bu78Quc46jOPfh02qaWux2CCapE2KURvF2pKGarEXqRctSnuh0hYVLIJQrI3iv+qleCMVqaVBTGuFYrRIjFJE2pQogiixMYm2SSyEGJO1Kd1cxFqRbnbz82LO4iRudGf/nZzd7weGOeedObO/4WV3n3nPO+eV9D+dL+HxbBNhcnT8NKfGJu6L0fFxRsd6j/0nIJ7+74A41YA5fprRseJ1M7h492wwtEmSpE46M0wufH4XWZIkqQMMbZIkSR1gaJMkSeoAQ5skSVIHGNokSZI6wNAmSZLUAYY2SZKkDmgltCW5OcnzSQ4luW+Sx29MsifJWJIPn/XY5Ul2JDmY5ECS1fNVtyRJUlvmPbQlGQI2A7cA64A7k6w762lHgU8AP5zkJR4D7n4/P/8AAAWESURBVK+qq4ANwEtzV60kSdL5oY0VETYAh6rqMECSx4HbgAMTT6iqI81jp/sPbMLdkqp6unneK/NUsyRJUqvaOD26AjjWtz/ctE3FW4GTSbYn2Zvk/mbk7gxJPpVkd5LdIyMjs1CyJElSu9oIbZmkraZ47BLgfcAXgOuAK+mdRj3zxaq2VNX6qlq/fPny6dYpSZJ03mgjtA0Dq/r2VwIvDHDs3qo6XFVjwM+Aa2e5PkmSpPNOG6HtGWBtkjVJLgDuAJ4Y4NhLk0wMn22kby6cJEnSQjXvoa0ZIbsHeAo4CPy4qvYn+VqSWwGSXJdkGPgI8N0k+5tjx+mdGt2Z5Fl6p1ofnu/3IEmSNN9SNdXpZN2UZAT46zz8qMuAE/PwczQ37L/usw+7zz7sPvtw5q6oqkkn5C/40DZfkuyuqvVt16Hpsf+6zz7sPvuw++zDueUyVpIkSR1gaJMkSeoAQ9vs2dJ2AZoR+6/77MPusw+7zz6cQ85pkyRJ6gBH2iRJkjrA0CZJktQBhrYZSnJzkueTHEpyX9v1aDBJViX5VZKDSfYn2dR2TRpckqEke5P8vO1aNLgkr0+yLcmfm9/F97RdkwaT5HPN39DnkvwoyWvbrmkhMrTNQJIhYDNwC7AOuDPJunar0oDGgM9X1VXA9cCn7cNO2kRvhRV107eBJ6vq7cDV2JedkmQFcC+wvqreCQzRW6JSs8zQNjMbgEPNAvajwOPAbS3XpAFU1fGq2tNs/4PeP4sV7ValQSRZCXwAeKTtWjS4JJcANwLfA6iq0ao62W5VmoYlwLIkS4ALgRdarmdBMrTNzArgWN/+MP7D76wkq4FrgF3tVqIBPQB8CTjddiGaliuBEeD7zSnuR5Jc1HZRmrqq+hvwTeAocBz4e1XtaLeqhcnQNjOZpM1rqHRQkouBnwCfraqX265HU5Pkg8BLVfWHtmvRtC0BrgW+U1XXAP8EnB/cIUkupXeWaQ3wZuCiJB9rt6qFydA2M8PAqr79lTgk3DlJltILbFuranvb9WggNwC3JjlCb3rCxiQ/aLckDWgYGK6qiRHubfRCnLrj/cBfqmqkqk4B24H3tlzTgmRom5lngLVJ1iS5gN7EyydarkkDSBJ6c2kOVtW32q5Hg6mqr1TVyqpaTe/375dV5Sf8DqmqF4FjSd7WNN0EHGixJA3uKHB9kgubv6k34ZdJ5sSStgvosqoaS3IP8BS9b8s8WlX7Wy5Lg7kB+DjwbJJ9TdtXq+oXLdYkLTafAbY2H34PA3e1XI8GUFW7kmwD9tD7Rv5eXM5qTriMlSRJUgd4elSSJKkDDG2SJEkdYGiTJEnqAEObJElSBxjaJEmSOsDQJmlRSzKeZF/fbdauxp9kdZLnZuv1JC1uXqdN0mL3r6p6d9tFSNL/40ibJE0iyZEk30jy++b2lqb9iiQ7k/ypub+8aX9jkp8m+WNzm1jGZyjJw0n2J9mRZFlrb0pSpxnaJC12y846PXp732MvV9UG4CHggabtIeCxqnoXsBV4sGl/EPh1VV1Nb+3MidVR1gKbq+odwEngQ3P8fiQtUK6IIGlRS/JKVV08SfsRYGNVHU6yFHixqt6Q5ATwpqo61bQfr6rLkowAK6vq1b7XWA08XVVrm/0vA0ur6utz/84kLTSOtEnSudU5ts/1nMm82rc9jnOJJU2ToU2Szu32vvvfNdu/Be5otj8K/KbZ3gncDZBkKMkl81WkpMXBT3ySFrtlSfb17T9ZVROX/XhNkl30PuDe2bTdCzya5IvACHBX074J2JLkk/RG1O4Gjs959ZIWDee0SdIkmjlt66vqRNu1SBJ4elSSJKkTHGmTJEnqAEfaJEmSOsDQJkmS1AGGNkmSpA4wtEmSJHWAoU2SJKkD/g2vJV3HA6oIFQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 720x1080 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_history(hist7)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_unbeer_keras(df, userid):\n",
    "    beers = df[df['user']==userid]['beer'].unique().tolist() #해당 유저가 평가한 beer\n",
    "    total_beers = df['beer'].unique().tolist() #모든 beer\n",
    "    unbeers = [beer for beer in total_beers if beer not in beers] #해당 유저가 평가안한 beer\n",
    "    return unbeers\n",
    "\n",
    "def recomm_beer_by_keras(userid, model, top_n):\n",
    "    unbeers = get_unbeer_keras(df,userid)\n",
    "    predict =[v[0][0] for v in  [model.predict([[userid],[beer]]) for beer in unbeers]]\n",
    "    ndf= pd.DataFrame({'beer':unbeers,'rating':predict})\n",
    "    final = ndf.sort_values(by='rating', ascending=False)[:top_n]\n",
    "    return final\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Wall time: 3min 1s\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>beer</th>\n",
       "      <th>rating</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>227886</th>\n",
       "      <td>53386</td>\n",
       "      <td>5.018239</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>252789</th>\n",
       "      <td>101353</td>\n",
       "      <td>4.998735</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8159</th>\n",
       "      <td>91882</td>\n",
       "      <td>4.969511</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>270318</th>\n",
       "      <td>179053</td>\n",
       "      <td>4.969384</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>208315</th>\n",
       "      <td>135645</td>\n",
       "      <td>4.965288</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>221111</th>\n",
       "      <td>65312</td>\n",
       "      <td>4.959350</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>186538</th>\n",
       "      <td>68022</td>\n",
       "      <td>4.937458</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>184632</th>\n",
       "      <td>145113</td>\n",
       "      <td>4.932275</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>269898</th>\n",
       "      <td>102640</td>\n",
       "      <td>4.930546</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>43781</th>\n",
       "      <td>154466</td>\n",
       "      <td>4.930435</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          beer    rating\n",
       "227886   53386  5.018239\n",
       "252789  101353  4.998735\n",
       "8159     91882  4.969511\n",
       "270318  179053  4.969384\n",
       "208315  135645  4.965288\n",
       "221111   65312  4.959350\n",
       "186538   68022  4.937458\n",
       "184632  145113  4.932275\n",
       "269898  102640  4.930546\n",
       "43781   154466  4.930435"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "unbeers = get_unbeer_keras(df2, 190)\n",
    "recomm_beer_by_keras(190, re7model, 10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(309542, 50)"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "## get weights of the beers embedding matrix\n",
    "beer_embedding_weights = re7model.layers[3].get_weights()[0] #3은 beer-embedding : layer 확인!\n",
    "beer_embedding_weights.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(3, 309542)"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "## import PCA\n",
    "pca = PCA(n_components = 3) ## use 3 components\n",
    "beer_embedding_weights_t = np.transpose(beer_embedding_weights) ## pass the transpose of the embedding matrix\n",
    "beer_pca = pca.fit(beer_embedding_weights_t) ## fit\n",
    "\n",
    "## display the resulting matrix dimensions\n",
    "beer_pca.components_.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([9.9840343e-01, 1.1468965e-03, 5.3597993e-05], dtype=float32)"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "## display the variance explained by the 3 components\n",
    "beer_pca.explained_variance_ratio_\n",
    " #variance 가 아주 낮으면 좋은 해석을 기대하기 어렵다.\n",
    "    #첫번째 PC 분산은 약 얼마? 78%로, dropout 규제를 넣었는데도 성능이 좋다!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "from operator import itemgetter\n",
    "\n",
    "## extract first PCA\n",
    "pca0 = beer_pca.components_[0]\n",
    "\n",
    "## get the value (pca0, beer title)\n",
    "beer_comp0 = [(f, beers_dict[i]) for f,i in zip(pca0, list(beers_dict.keys()))]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[(0.01126184, 'Pilsner'),\n",
       " (0.011077536, 'Morning Juice 6'),\n",
       " (0.010897062, 'Sunday Driving'),\n",
       " (0.0103482315, 'Castalia IPA'),\n",
       " (0.010331211, 'Würzburger Hofbräu Premium Pilsner'),\n",
       " (0.0102425, 'Oktoberfest'),\n",
       " (0.010219849, 'Cream Ale'),\n",
       " (0.010095107, 'Dark Secret Barrel Aged Coffee Bean'),\n",
       " (0.009872162, 'Bean Me Up Espresso Milk Stout'),\n",
       " (0.009774746, 'Passion For Me')]"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "## beers corresponding to the highest values of pca0\n",
    "sorted(beer_comp0, key = itemgetter(0), reverse = True)[:10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[(-0.006448596, 'Left Coast Hop Juice'),\n",
       " (-0.0063365065, 'Double Stuffed Marshmallow and Chocolate Cranium'),\n",
       " (-0.006229367, 'Wunder Pils'),\n",
       " (-0.006184741, 'Barrel Aged Edition - Scotch'),\n",
       " (-0.0061689806, '423 Blonde Ale'),\n",
       " (-0.0061501786, 'Communication Breakdown'),\n",
       " (-0.006090207, 'Icarus Saison'),\n",
       " (-0.0060554505, 'Blauw'),\n",
       " (-0.0060453913, 'La Montagnaise'),\n",
       " (-0.0060432586, 'Snowman Make Like A Tree And Leaf Maple Porter')]"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "## beers corresponding to the lowest values of pca0\n",
    "sorted(beer_comp0, key = itemgetter(0))[:10]"
   ]
  },
  {
   "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": "3.7.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}