01_mln-baselines/goodreads/goodreads-descriptive-analysis.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "pycharm": {
     "name": "#%% md\n"
    }
   },
   "source": [
    "# Goodreads descriptive analysis\n",
    "\n",
    "- data source: https://sites.google.com/eng.ucsd.edu/ucsdbookgraph/home"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "outputs": [],
   "source": [
    "import gzip\n",
    "import json\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import os\n",
    "pd.options.display.float_format = '{:,}'.format\n",
    "\n",
    "import seaborn as sns\n",
    "sns.set_style(\"whitegrid\")\n",
    "sns.set(font_scale=1.8)\n",
    "\n",
    "RANDOM = 2021\n",
    "\n",
    "figure_path = '/home/weiss/git/thesis/doc/figures/'\n",
    "\n",
    "DIR = '/home/weiss/rs_data/goodreads/'"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "outputs": [],
   "source": [
    "def count_lines(file_name):\n",
    "    \"\"\"count number of lines in a given file\"\"\"\n",
    "    print('counting file:', file_name)\n",
    "    count = 0\n",
    "    with gzip.open(file_name) as fin:\n",
    "        for l in fin:\n",
    "            count += 1\n",
    "    print('done!')\n",
    "    return count"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "outputs": [],
   "source": [
    "#n_book = count_lines(os.path.join(DIR, 'goodreads_books.json.gz'))\n",
    "#n_work = count_lines(os.path.join(DIR, 'goodreads_book_works.json.gz'))\n",
    "#n_author = count_lines(os.path.join(DIR, 'goodreads_book_authors.json.gz'))\n",
    "#n_series = count_lines(os.path.join(DIR, 'goodreads_book_series.json.gz'))"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "outputs": [],
   "source": [
    "def load_data(file_name, head = 500):\n",
    "    \"\"\"Load a given GZIP compressed JSON file\"\"\"\n",
    "    count = 0\n",
    "    data = []\n",
    "    with gzip.open(file_name) as fin:\n",
    "        for l in fin:\n",
    "            d = json.loads(l)\n",
    "            count += 1\n",
    "            data.append(d)\n",
    "\n",
    "            # break if reaches the 100th line\n",
    "            if (head is not None) and (count > head):\n",
    "                break\n",
    "    return data"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "outputs": [],
   "source": [
    "#df_book_stats = pd.DataFrame([n_book, n_work, n_author, n_series], dtype = float,\n",
    "#                             columns = ['count'],\n",
    "#                             index = ['# book', '# work', '# author', '# series'])\n",
    "#print(df_book_stats)"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "outputs": [
    {
     "data": {
      "text/plain": "{'user_id': '8842281e1d1347389f2ab93d60773d4d',\n 'book_id': '897461',\n 'review_id': '91e06a861712ad5a654f7c9829423049',\n 'rating': 5,\n 'review_text': 'An epic book - loved it, and wish I had found it before now. It does start a bit slow, but then really picks up. It\\'s the story of a young man in 1905ish who graduates from Harvard and then is sent by his rich uncle to be the consul to the made up country of Islandia, which is closed to trade with the outside world. \\n Islandia is described so clearly you could almost believe it wasn\\'t a made up country. You can feel it\\'s beauty, it\\'s pureness, and it\\'s culure, coming through the pages. \\n Many of us went through idealistic phases in our 20\\'s, and John Lang is no exception. He is trying to decide what will make him happy: Islandia, or a life in New York as a businessman. He is also searching for love, and seems to have a crush on every Islandian woman he meets! \\n Wright does an amazing job of describing what the Islandian culture is like. They are largely farmers: they work in the field all day, and their land provides what they need to eat and money to buy the rest. They don\\'t have to work super hard, and have plenty of time for leisure. Their society closed and smaller, so people are very decent to each other. There aren\\'t very many towns or central social hubs, so socializing is often accomplished by staying as a guest in someones farm house as you are traveling through the country, and you can do that totally unannounced. \\n Islandia is a utopia of sorts - they believe that progress in their culture will not lead to improving their lives. It is an interesting question - does technological progress actually end up making our lives better? Technology is supposed to improve the quality of our lives. We have better medicines, cleaner cities, we can communicate instantly around the world, and travel on cars and trains and planes. \\n But we lack the notion of a \"home\" that spans across generations or even multiple family units. Cousins and uncles and grandparents are only seen several times a year during family parties and holidays. In a sense, our society has shifted from being family centric to friend centric, as friends are the people we hang out with socially in the cities, where most of us live. This gives us often a richer variety of people and ideas to interact with, but the the connections are often not as close, and that is sad.',\n 'date_added': 'Wed Feb 06 20:42:04 -0800 2008',\n 'date_updated': 'Wed Mar 22 11:46:00 -0700 2017',\n 'read_at': 'Mon Mar 07 00:00:00 -0800 2011',\n 'started_at': 'Wed Jan 26 00:00:00 -0800 2011',\n 'n_votes': 7,\n 'n_comments': 0}"
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "reviews = load_data(os.path.join(DIR, 'goodreads_reviews_dedup.json.gz'))\n",
    "np.random.choice(reviews)"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "outputs": [],
   "source": [
    "# the relevant file is 'goodreads_interactions.csv'\n",
    "\n",
    "PATH_IN = os.path.join(DIR, 'goodreads_interactions.csv')\n",
    "df_interactions = pd.read_csv(PATH_IN)"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "=== first 5 records ===\n",
      "   user_id  book_id  is_read  rating  is_reviewed\n",
      "0        0      948        1       5            0\n",
      "1        0      947        1       5            1\n",
      "2        0      946        1       5            0\n",
      "3        0      945        1       5            0\n",
      "4        0      944        1       5            0\n",
      "size 228648342\n",
      "unique users 876145\n",
      "unique books 2360650\n"
     ]
    }
   ],
   "source": [
    "print('=== first 5 records ===')\n",
    "print(df_interactions.head())\n",
    "print(f'size {df_interactions.rating.size}')\n",
    "print(f'unique users {df_interactions.user_id.unique().size}')\n",
    "print(f'unique books {df_interactions.book_id.unique().size}')"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "=== Any empty cells ===\n",
      "user_id        False\n",
      "book_id        False\n",
      "is_read        False\n",
      "rating         False\n",
      "is_reviewed    False\n",
      "dtype: bool\n"
     ]
    }
   ],
   "source": [
    "print('=== Any empty cells ===')\n",
    "print(df_interactions.isnull().any())"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "=== duplicated records ===\n",
      "Empty DataFrame\n",
      "Columns: [user_id, book_id, is_read, rating, is_reviewed]\n",
      "Index: []\n"
     ]
    }
   ],
   "source": [
    "print('=== duplicated records ===')\n",
    "print(df_interactions[df_interactions.duplicated(['user_id', 'book_id'], keep=False)])"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "user_id\n",
      "165733    119777\n",
      "339532    101295\n",
      "59545      82114\n",
      "311008     79168\n",
      "396456     77057\n",
      "           ...  \n",
      "694442         1\n",
      "694429         1\n",
      "694424         1\n",
      "694422         1\n",
      "828859         1\n",
      "Length: 876145, dtype: int64\n"
     ]
    },
    {
     "data": {
      "text/plain": "<Figure size 432x288 with 1 Axes>",
      "image/png": "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\n"
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "rating_count_by_user = df_interactions.groupby('user_id').size().sort_values(ascending=False)\n",
    "print(rating_count_by_user)  # contains unrealistic users (way too many ratings)\n",
    "size = rating_count_by_user.size\n",
    "\n",
    "df = pd.DataFrame(dict(sorted_user_id=range(1,size+1), rating_counts=rating_count_by_user.values))\n",
    "\n",
    "sns.set(style='whitegrid', font_scale=1.5, rc={'xtick.labelsize':12, 'ytick.labelsize':12})\n",
    "image = sns.relplot(data=df,\n",
    "                    x=\"sorted_user_id\",\n",
    "                    y=\"rating_counts\",\n",
    "                    kind=\"line\",\n",
    "                    height=4,\n",
    "                    aspect=1.5\n",
    "                    )\n",
    "\n",
    "image.set(xlabel=\"users sorted desc. by # ratings\",\n",
    "          ylabel=\"# ratings\"\n",
    "          )\n",
    "\n",
    "image.savefig(figure_path + '/goodreads-long-tail-distribution.png', dpi=300, bbox_inches='tight')"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/weiss/.conda/envs/GNN/lib/python3.7/site-packages/seaborn/categorical.py:3704: UserWarning: The `factorplot` function has been renamed to `catplot`. The original name will be removed in a future release. Please update your code. Note that the default `kind` in `factorplot` (`'point'`) has changed `'strip'` in `catplot`.\n",
      "  warnings.warn(msg)\n",
      "/home/weiss/.conda/envs/GNN/lib/python3.7/site-packages/seaborn/_decorators.py:43: FutureWarning: Pass the following variable as a keyword arg: x. From version 0.12, the only valid positional argument will be `data`, and passing other arguments without an explicit keyword will result in an error or misinterpretation.\n",
      "  FutureWarning\n"
     ]
    },
    {
     "data": {
      "text/plain": "<Figure size 360x360 with 1 Axes>",
      "image/png": "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\n"
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# rating distribution before cleanup\n",
    "\n",
    "sns.set(style='whitegrid', font_scale=1.3)\n",
    "g = sns.factorplot(\"rating\", data=df_interactions, kind='count')\n",
    "g.set_ylabels(\"Total number of ratings\")\n",
    "g.savefig(figure_path + '/goodreads-rating-distribution-with-null.png', dpi=300, bbox_inches='tight')"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "=== first 5 records ===\n",
      "   user_id  book_id  is_read  rating  is_reviewed\n",
      "0        0      948        1       5            0\n",
      "1        0      947        1       5            1\n",
      "2        0      946        1       5            0\n",
      "3        0      945        1       5            0\n",
      "4        0      944        1       5            0\n",
      "# ratings 228648342\n",
      "# users 876145\n",
      "# books 2360650\n"
     ]
    }
   ],
   "source": [
    "print('=== first 5 records ===')\n",
    "print(df_interactions.head())\n",
    "print(f'# ratings {df_interactions.rating.size}')\n",
    "print(f'# users {df_interactions.user_id.unique().size}')\n",
    "print(f'# books {df_interactions.book_id.unique().size}')"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "outputs": [
    {
     "data": {
      "text/plain": "           user_id  book_id  is_read  rating  is_reviewed\n0                0      948        1       5            0\n1                0      947        1       5            1\n2                0      946        1       5            0\n3                0      945        1       5            0\n4                0      944        1       5            0\n...            ...      ...      ...     ...          ...\n228648336   876144   367428        1       3            0\n228648338   876144    23847        1       4            0\n228648339   876144    23950        1       3            0\n228648340   876144   374106        1       5            1\n228648341   876144   351607        1       5            1\n\n[104551549 rows x 5 columns]",
      "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_id</th>\n      <th>book_id</th>\n      <th>is_read</th>\n      <th>rating</th>\n      <th>is_reviewed</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>0</th>\n      <td>0</td>\n      <td>948</td>\n      <td>1</td>\n      <td>5</td>\n      <td>0</td>\n    </tr>\n    <tr>\n      <th>1</th>\n      <td>0</td>\n      <td>947</td>\n      <td>1</td>\n      <td>5</td>\n      <td>1</td>\n    </tr>\n    <tr>\n      <th>2</th>\n      <td>0</td>\n      <td>946</td>\n      <td>1</td>\n      <td>5</td>\n      <td>0</td>\n    </tr>\n    <tr>\n      <th>3</th>\n      <td>0</td>\n      <td>945</td>\n      <td>1</td>\n      <td>5</td>\n      <td>0</td>\n    </tr>\n    <tr>\n      <th>4</th>\n      <td>0</td>\n      <td>944</td>\n      <td>1</td>\n      <td>5</td>\n      <td>0</td>\n    </tr>\n    <tr>\n      <th>...</th>\n      <td>...</td>\n      <td>...</td>\n      <td>...</td>\n      <td>...</td>\n      <td>...</td>\n    </tr>\n    <tr>\n      <th>228648336</th>\n      <td>876144</td>\n      <td>367428</td>\n      <td>1</td>\n      <td>3</td>\n      <td>0</td>\n    </tr>\n    <tr>\n      <th>228648338</th>\n      <td>876144</td>\n      <td>23847</td>\n      <td>1</td>\n      <td>4</td>\n      <td>0</td>\n    </tr>\n    <tr>\n      <th>228648339</th>\n      <td>876144</td>\n      <td>23950</td>\n      <td>1</td>\n      <td>3</td>\n      <td>0</td>\n    </tr>\n    <tr>\n      <th>228648340</th>\n      <td>876144</td>\n      <td>374106</td>\n      <td>1</td>\n      <td>5</td>\n      <td>1</td>\n    </tr>\n    <tr>\n      <th>228648341</th>\n      <td>876144</td>\n      <td>351607</td>\n      <td>1</td>\n      <td>5</td>\n      <td>1</td>\n    </tr>\n  </tbody>\n</table>\n<p>104551549 rows × 5 columns</p>\n</div>"
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# remove ratings with value 0\n",
    "df_interactions= df_interactions[df_interactions['rating'] != 0]\n",
    "df_interactions"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/weiss/.conda/envs/GNN/lib/python3.7/site-packages/seaborn/categorical.py:3704: UserWarning: The `factorplot` function has been renamed to `catplot`. The original name will be removed in a future release. Please update your code. Note that the default `kind` in `factorplot` (`'point'`) has changed `'strip'` in `catplot`.\n",
      "  warnings.warn(msg)\n",
      "/home/weiss/.conda/envs/GNN/lib/python3.7/site-packages/seaborn/_decorators.py:43: FutureWarning: Pass the following variable as a keyword arg: x. From version 0.12, the only valid positional argument will be `data`, and passing other arguments without an explicit keyword will result in an error or misinterpretation.\n",
      "  FutureWarning\n"
     ]
    },
    {
     "data": {
      "text/plain": "<Figure size 360x360 with 1 Axes>",
      "image/png": "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\n"
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# rating distribution after cleanup\n",
    "sns.set(style='whitegrid', font_scale=1.3)\n",
    "g = sns.factorplot(\"rating\", data=df_interactions, kind='count')\n",
    "g.set_ylabels(\"# ratings\")\n",
    "g.savefig(figure_path + '/goodreads-rating-distribution.png', dpi=300, bbox_inches='tight')"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "True     424154\n",
      "False    392217\n",
      "Name: user_id, dtype: int64\n"
     ]
    }
   ],
   "source": [
    "# Keep entries where the user has rated more than n items and less than m items\n",
    "\n",
    "n = 50\n",
    "m = 1000\n",
    "\n",
    "counts = df_interactions['user_id'].value_counts()\n",
    "mask = (counts >= n) & (counts <= m)\n",
    "print(mask.value_counts())\n",
    "df_interactions_cut = df_interactions[df_interactions['user_id'].isin(mask[mask == True].index)]"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "=== first 5 records ===\n",
      "   user_id  book_id  is_read  rating  is_reviewed\n",
      "0        0      948        1       5            0\n",
      "1        0      947        1       5            1\n",
      "2        0      946        1       5            0\n",
      "3        0      945        1       5            0\n",
      "4        0      944        1       5            0\n",
      "# size 81569320\n",
      "# users 424154\n",
      "# books 2116340\n"
     ]
    }
   ],
   "source": [
    "print('=== first 5 records ===')\n",
    "print(df_interactions_cut.head())\n",
    "print(f'# size {df_interactions_cut[\"rating\"].size}')\n",
    "print(f'# users {df_interactions_cut[\"user_id\"].unique().size}')\n",
    "print(f'# books {df_interactions_cut.book_id.unique().size}')"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "user_id\n",
      "222640    1000\n",
      "67945     1000\n",
      "128540    1000\n",
      "20886     1000\n",
      "362086    1000\n",
      "          ... \n",
      "482643      50\n",
      "482646      50\n",
      "54187       50\n",
      "604658      50\n",
      "716827      50\n",
      "Length: 424154, dtype: int64\n",
      "<class 'pandas.core.series.Series'>\n",
      "424154\n"
     ]
    },
    {
     "data": {
      "text/plain": "<Figure size 360x360 with 1 Axes>",
      "image/png": "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\n"
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# power law?\n",
    "rating_count_by_user = df_interactions_cut.groupby('user_id').size().sort_values(ascending=False)\n",
    "print(rating_count_by_user)\n",
    "print(type(rating_count_by_user))\n",
    "size = rating_count_by_user.size\n",
    "print(size)\n",
    "\n",
    "df = pd.DataFrame(dict(sorted_user_id=range(1,size+1), rating_counts=rating_count_by_user.values))\n",
    "\n",
    "sns.set(style='whitegrid', font_scale=1.3, rc={'xtick.labelsize':12, 'ytick.labelsize':12})\n",
    "image = sns.relplot(data=df,\n",
    "                    x=\"sorted_user_id\",\n",
    "                    y=\"rating_counts\",\n",
    "                    kind=\"line\",\n",
    "                    )\n",
    "\n",
    "image.set(xlabel=\"users sorted desc. by # ratings\",\n",
    "          ylabel=\"# ratings\"\n",
    "          )\n",
    "\n",
    "\n",
    "image.savefig(figure_path + '/goodreads-long-tail-distribution-cut.png', dpi=300, bbox_inches='tight')"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "outputs": [
    {
     "data": {
      "text/plain": "3.946898061673188"
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_interactions_cut['rating'].mean()"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "outputs": [
    {
     "data": {
      "text/plain": "4.0"
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_interactions_cut['rating'].median()"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "outputs": [],
   "source": [],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%"
    }
   }
  }
 ],
 "metadata": {
  "kernelspec": {
   "name": "pycharm-f57757a1",
   "language": "python",
   "display_name": "PyCharm (rs-via-gnn)"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}