notebooks/dataset_20ng.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Download and pre-process the _20 NewsGroup_ dataset."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "pycharm": {
     "is_executing": false,
     "name": "#%%\n"
    }
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[nltk_data] Downloading package stopwords to\n",
      "[nltk_data]     /Users/pasquale/nltk_data...\n",
      "[nltk_data]   Package stopwords is already up-to-date!\n",
      "[nltk_data] Downloading package wordnet to\n",
      "[nltk_data]     /Users/pasquale/nltk_data...\n",
      "[nltk_data]   Package wordnet is already up-to-date!\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "True"
      ]
     },
     "execution_count": 1,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import re\n",
    "import pandas as pd\n",
    "import nltk\n",
    "nltk.download('stopwords')\n",
    "nltk.download('wordnet')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "pycharm": {
     "is_executing": false,
     "name": "#%%\n"
    }
   },
   "outputs": [],
   "source": [
    "data = pd.read_json('https://raw.githubusercontent.com/selva86/datasets/master/newsgroups.json')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "pycharm": {
     "is_executing": false,
     "name": "#%%\n"
    }
   },
   "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>content</th>\n",
       "      <th>target</th>\n",
       "      <th>target_names</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>From: lerxst@wam.umd.edu (where's my thing)\\nS...</td>\n",
       "      <td>7</td>\n",
       "      <td>rec.autos</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>From: guykuo@carson.u.washington.edu (Guy Kuo)...</td>\n",
       "      <td>4</td>\n",
       "      <td>comp.sys.mac.hardware</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>From: twillis@ec.ecn.purdue.edu (Thomas E Will...</td>\n",
       "      <td>4</td>\n",
       "      <td>comp.sys.mac.hardware</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>From: jgreen@amber (Joe Green)\\nSubject: Re: W...</td>\n",
       "      <td>1</td>\n",
       "      <td>comp.graphics</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>From: jcm@head-cfa.harvard.edu (Jonathan McDow...</td>\n",
       "      <td>14</td>\n",
       "      <td>sci.space</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                             content  target  \\\n",
       "0  From: lerxst@wam.umd.edu (where's my thing)\\nS...       7   \n",
       "1  From: guykuo@carson.u.washington.edu (Guy Kuo)...       4   \n",
       "2  From: twillis@ec.ecn.purdue.edu (Thomas E Will...       4   \n",
       "3  From: jgreen@amber (Joe Green)\\nSubject: Re: W...       1   \n",
       "4  From: jcm@head-cfa.harvard.edu (Jonathan McDow...      14   \n",
       "\n",
       "            target_names  \n",
       "0              rec.autos  \n",
       "1  comp.sys.mac.hardware  \n",
       "2  comp.sys.mac.hardware  \n",
       "3          comp.graphics  \n",
       "4              sci.space  "
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Subject: Re: 68LC040 vs. 68RC040 in Centris 650\n",
      "From: Bruce@hoult.actrix.gen.nz (Bruce Hoult)\n",
      "Lines: 16\n",
      "\n",
      "Zack T. Smith writes:\n",
      "> konpej@eua.ericsson.se (Per Ejeklint) writes:\n",
      "> \n",
      "> >Hm, maybe I'm missing something, but the Centris 650 has the '040 with FPU.\n",
      "> >At least the ones shipped here in Europe.\n",
      "> \n",
      "> You are indeed. The 4/80 model (shipped here) definitely does not have the\n",
      "> FPU. I own one; I know.\n",
      "\n",
      "\n",
      "No, he's not missing anything.  You're right that some models of the 650 ship in\n",
      "the USA without FPU or Ethernet.  Per Ejeklint is also right -- *all*, I repeat,\n",
      "*ALL* Centris 650's sold here in New Zealand and, I assume, Europe have the FPU\n",
      "and Ethernet.\n",
      "\n",
      "I know.  I bought a 650 4/80 and it has both FPU and Ethernet.\n",
      "\n"
     ]
    }
   ],
   "source": [
    "for x in data.loc[2186].content.split('\\n'):\n",
    "    print(x)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "pycharm": {
     "is_executing": false,
     "name": "#%%\n"
    }
   },
   "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>content</th>\n",
       "      <th>target</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>target_names</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>alt.atheism</th>\n",
       "      <td>480</td>\n",
       "      <td>480</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>comp.graphics</th>\n",
       "      <td>584</td>\n",
       "      <td>584</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>comp.os.ms-windows.misc</th>\n",
       "      <td>591</td>\n",
       "      <td>591</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>comp.sys.ibm.pc.hardware</th>\n",
       "      <td>590</td>\n",
       "      <td>590</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>comp.sys.mac.hardware</th>\n",
       "      <td>578</td>\n",
       "      <td>578</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>comp.windows.x</th>\n",
       "      <td>593</td>\n",
       "      <td>593</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>misc.forsale</th>\n",
       "      <td>585</td>\n",
       "      <td>585</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>rec.autos</th>\n",
       "      <td>594</td>\n",
       "      <td>594</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>rec.motorcycles</th>\n",
       "      <td>598</td>\n",
       "      <td>598</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>rec.sport.baseball</th>\n",
       "      <td>597</td>\n",
       "      <td>597</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>rec.sport.hockey</th>\n",
       "      <td>600</td>\n",
       "      <td>600</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>sci.crypt</th>\n",
       "      <td>595</td>\n",
       "      <td>595</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>sci.electronics</th>\n",
       "      <td>591</td>\n",
       "      <td>591</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>sci.med</th>\n",
       "      <td>594</td>\n",
       "      <td>594</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>sci.space</th>\n",
       "      <td>593</td>\n",
       "      <td>593</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>soc.religion.christian</th>\n",
       "      <td>599</td>\n",
       "      <td>599</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>talk.politics.guns</th>\n",
       "      <td>546</td>\n",
       "      <td>546</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>talk.politics.mideast</th>\n",
       "      <td>564</td>\n",
       "      <td>564</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>talk.politics.misc</th>\n",
       "      <td>465</td>\n",
       "      <td>465</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>talk.religion.misc</th>\n",
       "      <td>377</td>\n",
       "      <td>377</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                          content  target\n",
       "target_names                             \n",
       "alt.atheism                   480     480\n",
       "comp.graphics                 584     584\n",
       "comp.os.ms-windows.misc       591     591\n",
       "comp.sys.ibm.pc.hardware      590     590\n",
       "comp.sys.mac.hardware         578     578\n",
       "comp.windows.x                593     593\n",
       "misc.forsale                  585     585\n",
       "rec.autos                     594     594\n",
       "rec.motorcycles               598     598\n",
       "rec.sport.baseball            597     597\n",
       "rec.sport.hockey              600     600\n",
       "sci.crypt                     595     595\n",
       "sci.electronics               591     591\n",
       "sci.med                       594     594\n",
       "sci.space                     593     593\n",
       "soc.religion.christian        599     599\n",
       "talk.politics.guns            546     546\n",
       "talk.politics.mideast         564     564\n",
       "talk.politics.misc            465     465\n",
       "talk.religion.misc            377     377"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.groupby('target_names').count()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The dataset appear somehow balanced among classes."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "pycharm": {
     "is_executing": false
    }
   },
   "outputs": [],
   "source": [
    "lem = nltk.stem.WordNetLemmatizer()\n",
    "email_regex = r'([a-z0-9_-]+\\.)*[a-z0-9_-]+@[a-z0-9_-]+(\\.[a-z0-9_-]+)*\\.[a-z]{2,6}'\n",
    "\n",
    "def prepare_text(text):\n",
    "    text = text.lower()\n",
    "    text = re.sub('from:.+\\n', '', text)\n",
    "    text = re.sub('article-i\\.d\\.:.+\\n', '', text)\n",
    "    text = re.sub('nntp-posting-host:.+\\n', '', text)\n",
    "    text = re.sub('organization:.+\\n', '', text)\n",
    "    text = re.sub('x-newsreader:.+\\n', '', text)\n",
    "    text = re.sub('distribution:.+\\n', '', text)\n",
    "    text = re.sub('reply-to:.+\\n', '', text)\n",
    "    text = re.sub('from article <.+>.+\\n', '', text)\n",
    "    text = re.sub('in article .+\\n?.+\\n', '', text)\n",
    "    text = re.sub(r'\\d+', '', text)\n",
    "    text = text.split('=======')[0]\n",
    "    text = re.sub('.+ writes:?[\\s.]*\\n', '', text)\n",
    "    text = re.sub('<.+>', '', text)\n",
    "    text = re.sub(email_regex, '', text)\n",
    "    text = re.sub('[^\\s]+@[^\\s]+.+\\n', '', text)\n",
    "\n",
    "    text = nltk.tokenize.RegexpTokenizer(r'\\w+').tokenize(text)\n",
    "    text = [w for w in text if w != 'lines']\n",
    "    text = [lem.lemmatize(w) for w in text]\n",
    "    text = [w for w in text \n",
    "            if not(w in ['from', 'subject', 'edu', 'use', 'ke', 'one', 'would'] \n",
    "                   or w in nltk.corpus.stopwords.words('english')) \n",
    "            and '_' not in w]\n",
    "    text = [w for w in text if len(w)>=3]\n",
    "    return text"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "pycharm": {
     "is_executing": true
    }
   },
   "outputs": [],
   "source": [
    "data['content'] = data.content.apply(lambda x: prepare_text(x))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "pycharm": {
     "is_executing": true
    }
   },
   "outputs": [],
   "source": [
    "with open('../data/20ng.txt', 'w') as f:\n",
    "     for d in data['content']:\n",
    "            f.write(' '.join(d))\n",
    "            f.write('\\n')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "pycharm": {
     "is_executing": true
    }
   },
   "outputs": [],
   "source": [
    "with open('../data/20ng_labels.txt', 'w') as f:\n",
    "     for d in data['target_names']:\n",
    "            f.write(d)\n",
    "            f.write('\\n')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "pycharm": {
     "is_executing": true,
     "name": "#%% md\n"
    }
   },
   "source": [
    "## Some stats"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "pycharm": {
     "is_executing": true,
     "name": "#%%\n"
    }
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import gensim\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_words = data['content'].values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "pycharm": {
     "is_executing": true,
     "name": "#%%\n"
    }
   },
   "outputs": [],
   "source": [
    "id2word = gensim.corpora.Dictionary(data_words)\n",
    "\n",
    "corpus = [id2word.doc2bow(text) for text in data_words]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "pycharm": {
     "name": "#%% md \n"
    }
   },
   "source": [
    "Dimension of the dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "pycharm": {
     "is_executing": true,
     "name": "#%%\n"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "11314"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(data_words)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Average number of word per document"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "122.21681103058158"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.mean([len(d) for d in data_words])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Distribution of first 50 words"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "token_tf = [(id2word[token_id],freq) for token_id, freq in id2word.cfs.items()]\n",
    "token_tf.sort(key=lambda x: x[1], reverse = True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(figsize = (15,10))\n",
    "\n",
    "plt.bar(*zip(*token_tf[:50]))\n",
    "\n",
    "ax.tick_params(axis='x', rotation=70)\n",
    "ax.spines['right'].set_visible(False)\n",
    "ax.spines['top'].set_visible(False)\n",
    "\n",
    "plt.savefig('20ng_word_distribution.pdf')  "
   ]
  },
  {
   "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": "3.7.7"
  },
  "pycharm": {
   "stem_cell": {
    "cell_type": "raw",
    "metadata": {
     "collapsed": false
    },
    "source": []
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}