Scikit - 06 Text Processing.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd # Used for dataframe functions\n",
    "import json # parse json string\n",
    "import nltk # Natural language toolkit for TDIDF etc.\n",
    "from bs4 import BeautifulSoup # Parse html string .. to extract text\n",
    "import re # Regex parser \n",
    "import numpy as np # Linear algebbra \n",
    "from sklearn import * # machine learning\n",
    "import matplotlib.pyplot as plt # Visualization\n",
    "from wordcloud import WordCloud # Word cloud visualization\n",
    "import scipy #Sparse matrix \n",
    "\n",
    "np.set_printoptions(precision=4)\n",
    "pd.options.display.max_columns = 1000\n",
    "pd.options.display.max_rows = 10\n",
    "pd.options.display.float_format = lambda f: \"%.4f\" % f\n",
    "\n",
    "%matplotlib inline "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Run the following line when you run this notebook first time."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# nltk.download(\"punkt\")\n",
    "# nltk.download(\"stopwords\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3.3\n"
     ]
    }
   ],
   "source": [
    "print(nltk.__version__)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{\"label\":\"test\",\"sentiment\":\"pos\",\"name\":\"0_10.txt\",\"content\":\"I went and saw this movie last night after being coaxed to by a few friends of mine. I'll admit that I was reluctant to see it because from what I knew of Ashton Kutcher he was only able to do comedy. I was wrong. Kutcher played the character of Jake Fischer very well, and Kevin Costner played Ben Randall with such professionalism. The sign of a good movie is that it can toy with our emotions. This one did exactly that. The entire theater (which was sold out) was overcome by laughter during the first half of the movie, and were moved to tears during the second half. While exiting the theater I not only saw many women in tears, but many full grown men as well, trying desperately not to let anyone see them crying. This movie was great, and I suggest that you go see it before you judge.\"}\r\n"
     ]
    }
   ],
   "source": [
    "# The following line does not work on Windows system\n",
    "!head -n 1 /data/imdb-comments.json"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "data = []\n",
    "with open(\"/data/imdb-comments.json\", \"r\", encoding=\"utf8\") as f:\n",
    "    for l in f.readlines():\n",
    "        data.append(json.loads(l))    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "scrolled": true
   },
   "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>label</th>\n",
       "      <th>name</th>\n",
       "      <th>sentiment</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>26298</th>\n",
       "      <td>when i saw the movie at first i thought that i...</td>\n",
       "      <td>train</td>\n",
       "      <td>11169_7.txt</td>\n",
       "      <td>pos</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29718</th>\n",
       "      <td>Enjoyed viewing this film on TCM and watching ...</td>\n",
       "      <td>train</td>\n",
       "      <td>2998_7.txt</td>\n",
       "      <td>pos</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29482</th>\n",
       "      <td>\"Bend It Like Beckham\" reminds me of the best ...</td>\n",
       "      <td>train</td>\n",
       "      <td>2785_7.txt</td>\n",
       "      <td>pos</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>47234</th>\n",
       "      <td>This was one of the most ridiculous and badly ...</td>\n",
       "      <td>train</td>\n",
       "      <td>7511_1.txt</td>\n",
       "      <td>neg</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>31726</th>\n",
       "      <td>I have watched this movie on DVD a couple of t...</td>\n",
       "      <td>train</td>\n",
       "      <td>4804_7.txt</td>\n",
       "      <td>pos</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19482</th>\n",
       "      <td>Well its ten year's on since this film was rel...</td>\n",
       "      <td>test</td>\n",
       "      <td>5034_1.txt</td>\n",
       "      <td>neg</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16486</th>\n",
       "      <td>I get the feeling a lot of people liked this m...</td>\n",
       "      <td>test</td>\n",
       "      <td>2338_1.txt</td>\n",
       "      <td>neg</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>34495</th>\n",
       "      <td>I truly enjoyed this film. The acting was terr...</td>\n",
       "      <td>train</td>\n",
       "      <td>7297_10.txt</td>\n",
       "      <td>pos</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>31402</th>\n",
       "      <td>I saw this movie today and I have to say, it w...</td>\n",
       "      <td>train</td>\n",
       "      <td>4512_9.txt</td>\n",
       "      <td>pos</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>38008</th>\n",
       "      <td>I went into this film with expectations, from ...</td>\n",
       "      <td>train</td>\n",
       "      <td>10458_2.txt</td>\n",
       "      <td>neg</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                 content  label         name  \\\n",
       "26298  when i saw the movie at first i thought that i...  train  11169_7.txt   \n",
       "29718  Enjoyed viewing this film on TCM and watching ...  train   2998_7.txt   \n",
       "29482  \"Bend It Like Beckham\" reminds me of the best ...  train   2785_7.txt   \n",
       "47234  This was one of the most ridiculous and badly ...  train   7511_1.txt   \n",
       "31726  I have watched this movie on DVD a couple of t...  train   4804_7.txt   \n",
       "19482  Well its ten year's on since this film was rel...   test   5034_1.txt   \n",
       "16486  I get the feeling a lot of people liked this m...   test   2338_1.txt   \n",
       "34495  I truly enjoyed this film. The acting was terr...  train  7297_10.txt   \n",
       "31402  I saw this movie today and I have to say, it w...  train   4512_9.txt   \n",
       "38008  I went into this film with expectations, from ...  train  10458_2.txt   \n",
       "\n",
       "      sentiment  \n",
       "26298       pos  \n",
       "29718       pos  \n",
       "29482       pos  \n",
       "47234       neg  \n",
       "31726       pos  \n",
       "19482       neg  \n",
       "16486       neg  \n",
       "34495       pos  \n",
       "31402       pos  \n",
       "38008       neg  "
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "comments = pd.DataFrame.from_dict(data)\n",
    "comments.sample(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 50000 entries, 0 to 49999\n",
      "Data columns (total 4 columns):\n",
      "content      50000 non-null object\n",
      "label        50000 non-null object\n",
      "name         50000 non-null object\n",
      "sentiment    50000 non-null object\n",
      "dtypes: object(4)\n",
      "memory usage: 1.5+ MB\n"
     ]
    }
   ],
   "source": [
    "comments.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "test     25000\n",
       "train    25000\n",
       "Name: label, dtype: int64"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "comments.label.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "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>sentiment</th>\n",
       "      <th>neg</th>\n",
       "      <th>pos</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>label</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>test</th>\n",
       "      <td>12500</td>\n",
       "      <td>12500</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>train</th>\n",
       "      <td>12500</td>\n",
       "      <td>12500</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "sentiment    neg    pos\n",
       "label                  \n",
       "test       12500  12500\n",
       "train      12500  12500"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "comments.groupby([\"label\", \"sentiment\"]).content.count().unstack()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'When we started watching this series on cable, I had no idea how addictive it would be. Even when you hate a character, you hold back because they are so beautifully developed, you can almost understand why they react to frustration, fear, greed or temptation the way they do. It\\'s almost as if the viewer is experiencing one of Christopher\\'s learning curves.<br /><br />I can\\'t understand why Adriana would put up with Christopher\\'s abuse of her, verbally, physically and emotionally, but I just have to read the newspaper to see how many women can and do tolerate such behavior. Carmella has a dream house, endless supply of expensive things, but I\\'m sure she would give it up for a loving and faithful husband - or maybe not. That\\'s why I watch.<br /><br />It doesn\\'t matter how many times you watch an episode, you can find something you missed the first five times. We even watch episodes out of sequence (watch season 1 on late night with commercials but all the language, A&E with language censored, reruns on the Movie Network) - whenever they\\'re on, we\\'re there. We\\'ve been totally spoiled now.<br /><br />I also love the Malaprop\\'s. \"An albacore around my neck\" is my favorite of Johnny Boy. When these jewels have entered our family vocabulary, it is a sign that I should get a life. I will when the series ends, and I have collected all the DVD\\'s, and put the collection in my will.'"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.random.seed(1)\n",
    "v = list(comments[\"content\"].sample(1))[0]\n",
    "v"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'when we started watching this series on cable i had no idea how addictive it would be even when you hate a character you hold back because they are so beautifully developed you can almost understand why they react to frustration fear greed or temptation the way they do it s almost as if the viewer is experiencing one of christopher s learning curves i can t understand why adriana would put up with christopher s abuse of her verbally physically and emotionally but i just have to read the newspaper to see how many women can and do tolerate such behavior carmella has a dream house endless supply of expensive things but i m sure she would give it up for a loving and faithful husband or maybe not that s why i watch it doesn t matter how many times you watch an episode you can find something you missed the first five times we even watch episodes out of sequence watch season 1 on late night with commercials but all the language a e with language censored reruns on the movie network whenever they re on we re there we ve been totally spoiled now i also love the malaprop s an albacore around my neck is my favorite of johnny boy when these jewels have entered our family vocabulary it is a sign that i should get a life i will when the series ends and i have collected all the dvd s and put the collection in my will '"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def preprocess(text):\n",
    "    \n",
    "    # Remove html tags\n",
    "    text = BeautifulSoup(text.lower(), \"html5lib\").text \n",
    "    \n",
    "    # Replace the occurrences of multiple consecutive whilespaces \n",
    "    # with a single space (\" \")\n",
    "    text = re.sub(r\"[\\W]+\", \" \", text)\n",
    "    return text\n",
    "\n",
    "preprocess(v)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 32.6 s, sys: 51.8 ms, total: 32.6 s\n",
      "Wall time: 32.6 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "# Apply the preprocessing logic to all comments\n",
    "comments[\"content\"] = comments[\"content\"].apply(preprocess)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "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>content</th>\n",
       "      <th>label</th>\n",
       "      <th>name</th>\n",
       "      <th>sentiment</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>25496</th>\n",
       "      <td>there are many people in our lives that we mee...</td>\n",
       "      <td>train</td>\n",
       "      <td>10447_10.txt</td>\n",
       "      <td>pos</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>41520</th>\n",
       "      <td>sixth escapade for freddy krueger in which he ...</td>\n",
       "      <td>train</td>\n",
       "      <td>2369_4.txt</td>\n",
       "      <td>neg</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>48525</th>\n",
       "      <td>the first few minutes of the bodyguard do have...</td>\n",
       "      <td>train</td>\n",
       "      <td>8674_2.txt</td>\n",
       "      <td>neg</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26037</th>\n",
       "      <td>after reviewing this intense martial arts movi...</td>\n",
       "      <td>train</td>\n",
       "      <td>10934_10.txt</td>\n",
       "      <td>pos</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>43823</th>\n",
       "      <td>triumph of love is proof that not every coméd...</td>\n",
       "      <td>train</td>\n",
       "      <td>4441_3.txt</td>\n",
       "      <td>neg</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>39666</th>\n",
       "      <td>this is without a doubt the worst movie i have...</td>\n",
       "      <td>train</td>\n",
       "      <td>11950_2.txt</td>\n",
       "      <td>neg</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>40882</th>\n",
       "      <td>i had the opportunity to preview this film as ...</td>\n",
       "      <td>train</td>\n",
       "      <td>1795_2.txt</td>\n",
       "      <td>neg</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>43156</th>\n",
       "      <td>roeg has done some great movies but this a tur...</td>\n",
       "      <td>train</td>\n",
       "      <td>3841_1.txt</td>\n",
       "      <td>neg</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>37185</th>\n",
       "      <td>possible spoiler in some way how to alienate f...</td>\n",
       "      <td>train</td>\n",
       "      <td>9718_7.txt</td>\n",
       "      <td>pos</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>40330</th>\n",
       "      <td>sogo ishii can be a skilled filmmaker under th...</td>\n",
       "      <td>train</td>\n",
       "      <td>1298_3.txt</td>\n",
       "      <td>neg</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                 content  label          name  \\\n",
       "25496  there are many people in our lives that we mee...  train  10447_10.txt   \n",
       "41520  sixth escapade for freddy krueger in which he ...  train    2369_4.txt   \n",
       "48525  the first few minutes of the bodyguard do have...  train    8674_2.txt   \n",
       "26037  after reviewing this intense martial arts movi...  train  10934_10.txt   \n",
       "43823   triumph of love is proof that not every coméd...  train    4441_3.txt   \n",
       "39666  this is without a doubt the worst movie i have...  train   11950_2.txt   \n",
       "40882  i had the opportunity to preview this film as ...  train    1795_2.txt   \n",
       "43156  roeg has done some great movies but this a tur...  train    3841_1.txt   \n",
       "37185  possible spoiler in some way how to alienate f...  train    9718_7.txt   \n",
       "40330  sogo ishii can be a skilled filmmaker under th...  train    1298_3.txt   \n",
       "\n",
       "      sentiment  \n",
       "25496       pos  \n",
       "41520       neg  \n",
       "48525       neg  \n",
       "26037       pos  \n",
       "43823       neg  \n",
       "39666       neg  \n",
       "40882       neg  \n",
       "43156       neg  \n",
       "37185       pos  \n",
       "40330       neg  "
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "comments_train = comments[comments[\"label\"] == \"train\"]\n",
    "comments_train.sample(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "scrolled": true
   },
   "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>label</th>\n",
       "      <th>name</th>\n",
       "      <th>sentiment</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>21277</th>\n",
       "      <td>this is one of those movies where i was rootin...</td>\n",
       "      <td>test</td>\n",
       "      <td>6650_1.txt</td>\n",
       "      <td>neg</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12620</th>\n",
       "      <td>spoilersi m going to be as kind as i can about...</td>\n",
       "      <td>test</td>\n",
       "      <td>10108_1.txt</td>\n",
       "      <td>neg</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4485</th>\n",
       "      <td>as a former erasmus student i enjoyed this fil...</td>\n",
       "      <td>test</td>\n",
       "      <td>2788_9.txt</td>\n",
       "      <td>pos</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4744</th>\n",
       "      <td>francis ford coppola s masterpiece was a great...</td>\n",
       "      <td>test</td>\n",
       "      <td>301_10.txt</td>\n",
       "      <td>pos</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10329</th>\n",
       "      <td>some of the reviewers here have foolishly judg...</td>\n",
       "      <td>test</td>\n",
       "      <td>8047_10.txt</td>\n",
       "      <td>pos</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14366</th>\n",
       "      <td>this is by far one of the most pretentious fil...</td>\n",
       "      <td>test</td>\n",
       "      <td>11680_1.txt</td>\n",
       "      <td>neg</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1007</th>\n",
       "      <td>this is truly a funny movie his dance scene do...</td>\n",
       "      <td>test</td>\n",
       "      <td>10907_8.txt</td>\n",
       "      <td>pos</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4539</th>\n",
       "      <td>this film released in 1951 has the usual eleme...</td>\n",
       "      <td>test</td>\n",
       "      <td>2836_8.txt</td>\n",
       "      <td>pos</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7864</th>\n",
       "      <td>one of the best tv shows out there if not the ...</td>\n",
       "      <td>test</td>\n",
       "      <td>5829_10.txt</td>\n",
       "      <td>pos</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19582</th>\n",
       "      <td>this movie was jerry bruckheimer s idea to sel...</td>\n",
       "      <td>test</td>\n",
       "      <td>5124_4.txt</td>\n",
       "      <td>neg</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                 content label         name  \\\n",
       "21277  this is one of those movies where i was rootin...  test   6650_1.txt   \n",
       "12620  spoilersi m going to be as kind as i can about...  test  10108_1.txt   \n",
       "4485   as a former erasmus student i enjoyed this fil...  test   2788_9.txt   \n",
       "4744   francis ford coppola s masterpiece was a great...  test   301_10.txt   \n",
       "10329  some of the reviewers here have foolishly judg...  test  8047_10.txt   \n",
       "14366  this is by far one of the most pretentious fil...  test  11680_1.txt   \n",
       "1007   this is truly a funny movie his dance scene do...  test  10907_8.txt   \n",
       "4539   this film released in 1951 has the usual eleme...  test   2836_8.txt   \n",
       "7864   one of the best tv shows out there if not the ...  test  5829_10.txt   \n",
       "19582  this movie was jerry bruckheimer s idea to sel...  test   5124_4.txt   \n",
       "\n",
       "      sentiment  \n",
       "21277       neg  \n",
       "12620       neg  \n",
       "4485        pos  \n",
       "4744        pos  \n",
       "10329       pos  \n",
       "14366       neg  \n",
       "1007        pos  \n",
       "4539        pos  \n",
       "7864        pos  \n",
       "19582       neg  "
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "comments_test = comments[comments[\"label\"] == \"test\"]\n",
    "comments_test.sample(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train = comments_train[\"content\"].values\n",
    "y_train = np.where(comments_train.sentiment == \"pos\", 1, 0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_test = comments_test[\"content\"].values\n",
    "y_test = np.where(comments_test.sentiment == \"pos\", 1, 0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "('arabic', 'danish', 'dutch', 'english', 'finnish', 'french', 'german', 'hungarian', 'italian', 'norwegian', 'porter', 'portuguese', 'romanian', 'russian', 'spanish', 'swedish')\n"
     ]
    }
   ],
   "source": [
    "# http://snowball.tartarus.org/algorithms/porter/stemmer.html\n",
    "# http://www.nltk.org/howto/stem.html\n",
    "\n",
    "from nltk.stem.snowball import SnowballStemmer\n",
    "from nltk.stem.porter import PorterStemmer\n",
    "print(SnowballStemmer.languages)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "revival => reviv reviv revival\n",
      "allowance => allow allow allowance\n",
      "inference => infer infer inference\n",
      "relational => relat relat relational\n",
      "runner => runner runner runner\n",
      "runs => run run run\n",
      "ran => ran ran ran\n",
      "having => have have having\n",
      "generously => gener generous generously\n",
      "wasn't => wasn't wasn't wasn't\n",
      "leaves => leav leav leaf\n"
     ]
    }
   ],
   "source": [
    "porter = PorterStemmer()\n",
    "snowball = SnowballStemmer(\"english\")\n",
    "lemmatizer = nltk.wordnet.WordNetLemmatizer()\n",
    "\n",
    "for s in nltk.word_tokenize(\"\"\"\n",
    "            revival\n",
    "            allowance \n",
    "            inference \n",
    "            relational\n",
    "            runner\n",
    "            runs\n",
    "            ran\n",
    "            having\n",
    "            generously\n",
    "            wasn't\n",
    "            leaves\n",
    "            \"\"\"):\n",
    "    print(s, \"=>\", porter.stem(s)\n",
    "          , snowball.stem(s), lemmatizer.lemmatize(s, \"n\"))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "179 ['i', 'me', 'my', 'myself', 'we', 'our', 'ours', 'ourselves', 'you', \"you're\", \"you've\", \"you'll\", \"you'd\", 'your', 'yours', 'yourself', 'yourselves', 'he', 'him', 'his', 'himself', 'she', \"she's\", 'her', 'hers', 'herself', 'it', \"it's\", 'its', 'itself', 'they', 'them', 'their', 'theirs', 'themselves', 'what', 'which', 'who', 'whom', 'this', 'that', \"that'll\", 'these', 'those', 'am', 'is', 'are', 'was', 'were', 'be', 'been', 'being', 'have', 'has', 'had', 'having', 'do', 'does', 'did', 'doing', 'a', 'an', 'the', 'and', 'but', 'if', 'or', 'because', 'as', 'until', 'while', 'of', 'at', 'by', 'for', 'with', 'about', 'against', 'between', 'into', 'through', 'during', 'before', 'after', 'above', 'below', 'to', 'from', 'up', 'down', 'in', 'out', 'on', 'off', 'over', 'under', 'again', 'further', 'then', 'once', 'here', 'there', 'when', 'where', 'why', 'how', 'all', 'any', 'both', 'each', 'few', 'more', 'most', 'other', 'some', 'such', 'no', 'nor', 'not', 'only', 'own', 'same', 'so', 'than', 'too', 'very', 's', 't', 'can', 'will', 'just', 'don', \"don't\", 'should', \"should've\", 'now', 'd', 'll', 'm', 'o', 're', 've', 'y', 'ain', 'aren', \"aren't\", 'couldn', \"couldn't\", 'didn', \"didn't\", 'doesn', \"doesn't\", 'hadn', \"hadn't\", 'hasn', \"hasn't\", 'haven', \"haven't\", 'isn', \"isn't\", 'ma', 'mightn', \"mightn't\", 'mustn', \"mustn't\", 'needn', \"needn't\", 'shan', \"shan't\", 'shouldn', \"shouldn't\", 'wasn', \"wasn't\", 'weren', \"weren't\", 'won', \"won't\", 'wouldn', \"wouldn't\"]\n"
     ]
    }
   ],
   "source": [
    "stopwords = nltk.corpus.stopwords.words(\"english\")\n",
    "print(len(stopwords), stopwords)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['sun', 'shine', 'allow']"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#stemmer = SnowballStemmer(\"english\")\n",
    "stemmer = PorterStemmer()\n",
    "def my_tokenizer(s):\n",
    "    #terms = nltk.word_tokenize(s.lower())\n",
    "    terms = re.split(\"\\s\", s.lower())\n",
    "    terms = [re.sub(r\"[\\.!]\", \"\", v) for v in terms if len(v)>2]\n",
    "    terms = [v for v in terms if len(v)>2]\n",
    "    terms = [v for v in terms if v not in stopwords]\n",
    "    terms = [stemmer.stem(w) for w in terms]\n",
    "    terms = [term for term in terms if len(term) > 2]\n",
    "    return terms \n",
    "my_tokenizer(\"The sun is shining. Allowance wasn't!\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Vocabulary:  {'today': 10, 'wednesday': 14, 'today wednesday': 11, 'delhi': 2, 'weather': 12, 'hot': 5, 'delhi weather': 4, 'weather hot': 13, 'hot today': 6, 'road': 8, 'busi': 0, 'morn': 7, 'delhi road': 3, 'road busi': 9, 'busi morn': 1}\n"
     ]
    }
   ],
   "source": [
    "tfidf = feature_extraction.text.TfidfVectorizer(tokenizer=my_tokenizer\n",
    "                                                , ngram_range=(1, 2))\n",
    "\n",
    "corpus = [\"Today is Wednesday\"\n",
    "          , \"Delhi weather is hot today.\"\n",
    "          , \"Delhi roads are not busy in the morning\"]\n",
    "\n",
    "doc_term_matrix = tfidf.fit_transform(corpus)\n",
    "\n",
    "# returns term and index in the feature matrix\n",
    "print(\"Vocabulary: \", tfidf.vocabulary_) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "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>busi</th>\n",
       "      <th>busi morn</th>\n",
       "      <th>delhi</th>\n",
       "      <th>delhi road</th>\n",
       "      <th>delhi weather</th>\n",
       "      <th>hot</th>\n",
       "      <th>hot today</th>\n",
       "      <th>morn</th>\n",
       "      <th>road</th>\n",
       "      <th>road busi</th>\n",
       "      <th>today</th>\n",
       "      <th>today wednesday</th>\n",
       "      <th>weather</th>\n",
       "      <th>weather hot</th>\n",
       "      <th>wednesday</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.4736</td>\n",
       "      <td>0.6228</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.6228</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.3065</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.4030</td>\n",
       "      <td>0.4030</td>\n",
       "      <td>0.4030</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.3065</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.4030</td>\n",
       "      <td>0.4030</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.3899</td>\n",
       "      <td>0.3899</td>\n",
       "      <td>0.2965</td>\n",
       "      <td>0.3899</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.3899</td>\n",
       "      <td>0.3899</td>\n",
       "      <td>0.3899</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    busi  busi morn  delhi  delhi road  delhi weather    hot  hot today  \\\n",
       "0 0.0000     0.0000 0.0000      0.0000         0.0000 0.0000     0.0000   \n",
       "1 0.0000     0.0000 0.3065      0.0000         0.4030 0.4030     0.4030   \n",
       "2 0.3899     0.3899 0.2965      0.3899         0.0000 0.0000     0.0000   \n",
       "\n",
       "    morn   road  road busi  today  today wednesday  weather  weather hot  \\\n",
       "0 0.0000 0.0000     0.0000 0.4736           0.6228   0.0000       0.0000   \n",
       "1 0.0000 0.0000     0.0000 0.3065           0.0000   0.4030       0.4030   \n",
       "2 0.3899 0.3899     0.3899 0.0000           0.0000   0.0000       0.0000   \n",
       "\n",
       "   wednesday  \n",
       "0     0.6228  \n",
       "1     0.0000  \n",
       "2     0.0000  "
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "columns = [None] * len(tfidf.vocabulary_)\n",
    "for term in tfidf.vocabulary_:\n",
    "    columns[tfidf.vocabulary_[term]] = term\n",
    "columns\n",
    "scores = pd.DataFrame(doc_term_matrix.toarray()\n",
    "                      , columns= columns)\n",
    "scores"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 2min 18s, sys: 648 ms, total: 2min 19s\n",
      "Wall time: 2min 16s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "X_train_tfidf = tfidf.fit_transform(X_train)\n",
    "X_test_tfidf = tfidf.transform(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((25000, 1565606), (25000,), (25000, 1565606), (25000,))"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_test_tfidf.shape, y_test.shape, X_train_tfidf.shape, y_train.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's estimate the memory requirment if the data is presented in dense matrix format"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "145.80842107534409"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cell_count = np.product(X_train_tfidf.shape)\n",
    "bytes = cell_count * 4 \n",
    "GBs = bytes / (1024 ** 3) \n",
    "GBs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9998672943511969"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sparsity = 1 - X_train_tfidf.count_nonzero() / cell_count\n",
    "sparsity"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9998672943511969"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "1 - X_train_tfidf.nnz / cell_count"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Type of doc_term_matrix <class 'scipy.sparse.csr.csr_matrix'>\n"
     ]
    }
   ],
   "source": [
    "print(\"Type of doc_term_matrix\", type(X_train_tfidf))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Byte size of the training doc sparse doc"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.0386992022395134 GB\n"
     ]
    }
   ],
   "source": [
    "print(X_train_tfidf.data.nbytes / (1024.0 ** 3), \"GB\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Classification Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training accuracy:  0.93884\n",
      "Test accuracy:  0.8732\n",
      "CPU times: user 5 s, sys: 18.7 ms, total: 5.02 s\n",
      "Wall time: 4.27 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "lr = linear_model.LogisticRegression(C = 0.6, random_state = 1\n",
    "                            , n_jobs = 8, solver=\"saga\")\n",
    "lr.fit(X_train_tfidf, y_train)\n",
    "y_train_pred = lr.predict(X_train_tfidf)\n",
    "y_test_pred = lr.predict(X_test_tfidf)\n",
    "print(\"Training accuracy: \", metrics.accuracy_score(y_train, y_train_pred))\n",
    "print(\"Test accuracy: \", metrics.accuracy_score(y_test, y_test_pred))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fpr, tpr, thresholds = metrics.roc_curve(y_test,\n",
    "                        lr.predict_proba(X_test_tfidf)[:, [1]])\n",
    "auc = metrics.auc(fpr, tpr)\n",
    "\n",
    "plt.plot(fpr, tpr)\n",
    "plt.ylim(0, 1)\n",
    "plt.xlim(0, 1)\n",
    "plt.plot([0,1], [0,1], ls = \"--\", color = \"k\")\n",
    "plt.xlabel(\"False Postive Rate\")\n",
    "plt.ylabel(\"True Positive Rate\")\n",
    "plt.title(\"ROC Curve, auc: %.4f\" % auc);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "accuracy:  0.8584\n",
      "CPU times: user 605 ms, sys: 81.4 ms, total: 687 ms\n",
      "Wall time: 176 ms\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "from sklearn import naive_bayes, ensemble\n",
    "bayes = naive_bayes.MultinomialNB(alpha=1)\n",
    "bayes.fit(X_train_tfidf, y_train)\n",
    "print(\"accuracy: \", bayes.score(X_test_tfidf, y_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "accuracy:  0.7016\n",
      "CPU times: user 3min 14s, sys: 327 ms, total: 3min 14s\n",
      "Wall time: 3min 12s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "est = tree.DecisionTreeClassifier()\n",
    "est.fit(X_train_tfidf, y_train)\n",
    "print(\"accuracy: \", est.score(X_test_tfidf, y_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Top 50 terms:  ['bad', 'worst', 'wast', 'great', 'bore', 'terribl', 'love', 'noth', 'stupid', 'excel', 'poor', 'best', 'disappoint', 'beauti', 'wors', 'movi', 'poorli', 'good', 'perfect', 'fail', 'dull', 'well', 'tri', 'also', 'enjoy', 'act', 'look', 'pointless', 'one', 'ridicul', 'unfortun', 'save', 'one best', 'film', 'would', 'time', 'even', 'minut', 'perform', 'instead', 'see', 'much', 'plot', 'suppos', 'watch', 'mess', 'job', 'lame', 'show', 'annoy']\n"
     ]
    }
   ],
   "source": [
    "columns = [None] * len(tfidf.vocabulary_)\n",
    "for term in tfidf.vocabulary_:\n",
    "    columns[tfidf.vocabulary_[term]] = term\n",
    "result = pd.DataFrame({\"feature\": columns\n",
    "                    , \"importance\": est.feature_importances_})\n",
    "result = result.sort_values(\"importance\", ascending = False)\n",
    "result = result[result.importance > 0.0]\n",
    "print(\"Top 50 terms: \", list(result.feature[:50]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Important terms for a document "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "vocab_by_term = tfidf.vocabulary_\n",
    "vocab_by_idx = dict({(vocab_by_term[term], term) \n",
    "                                 for term in vocab_by_term})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\"{'bromwel': 180037, 'high': 644320, 'cartoon': 209092, 'comedi': 266965, 'ran': 1100682, 'time': 140\""
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "str(vocab_by_term)[:100]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\"{241553: 'christian distract', 1478662: 'victor manuel', 1563213: 'zarustica whole', 1009051: 'peac \""
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "str(vocab_by_idx)[:100]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Content:\n",
      " this isn t the comedic robin williams nor is it the quirky insane robin williams of recent thriller fame this is a hybrid of the classic drama without over dramatization mixed with robin s new love of the thriller but this isn t a thriller per se this is more a mystery suspense vehicle through which williams attempts to locate a sick boy and his keeper also starring sandra oh and rory culkin this suspense drama plays pretty much like a news report until william s character gets close to achieving his goal i must say that i was highly entertained though this movie fails to teach guide inspect or amuse it felt more like i was watching a guy williams as he was actually performing the actions from a third person perspective in other words it felt real and i was able to subscribe to the premise of the story all in all it s worth a watch though it s definitely not friday saturday night fare it rates a 7 7 10 from the fiend \n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "william         0.2223\n",
       "robin           0.1750\n",
       "robin william   0.1343\n",
       "thriller        0.1295\n",
       "dramat mix      0.1021\n",
       "                 ...  \n",
       "much            0.0221\n",
       "stori           0.0209\n",
       "charact         0.0206\n",
       "get             0.0198\n",
       "movi            0.0140\n",
       "Length: 162, dtype: float64"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "idx = 5\n",
    "print(\"Content:\\n\", X_train[idx])\n",
    "row = X_train_tfidf[idx]\n",
    "terms = [(vocab_by_idx[row.indices[i]], row.data[i])\n",
    "             for i, term in enumerate(row.indices)]\n",
    "pd.Series(dict(terms)).sort_values(ascending = False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "idx = 50\n",
    "row = X_train_tfidf[idx]\n",
    "terms = [(vocab_by_idx[row.indices[i]], row.data[i]) \n",
    "                 for i, term in enumerate(row.indices)]\n",
    "top_terms= list(pd.Series(dict(terms))\\\n",
    "                .sort_values(ascending = False)[:50].index)\n",
    "wc = WordCloud(background_color=\"white\", \n",
    "    width=500, height=500, max_words=50).generate(\"+\".join(top_terms))\n",
    "plt.figure(figsize=(10, 10))\n",
    "plt.imshow(wc)\n",
    "plt.axis(\"off\");"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Build Pipeline for classificaiton Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 1min 18s, sys: 448 ms, total: 1min 19s\n",
      "Wall time: 1min 15s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "tfidf =feature_extraction.text.TfidfVectorizer(\n",
    "              tokenizer=my_tokenizer\n",
    "            , stop_words = stopwords\n",
    "            , ngram_range=(1, 2)\n",
    "        )\n",
    "\n",
    "pipe = pipeline.Pipeline([\n",
    "    (\"tfidf\", tfidf),\n",
    "    (\"est\", linear_model.LogisticRegression(C = 1.0, random_state = 1\n",
    "                            , n_jobs = 8, solver=\"saga\"))\n",
    "])\n",
    "pipe.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pickle"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open(\"/tmp/model.pkl\", \"wb\") as f:\n",
    "    pickle.dump(pipe, f)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-rw-r--r--  1 abasar  wheel   123M Nov 23 06:59 /tmp/model.pkl\r\n"
     ]
    }
   ],
   "source": [
    "!ls -lh /tmp/model.pkl"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open(\"/tmp/model.pkl\", \"rb\") as f:\n",
    "    model = pickle.load(f)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [],
   "source": [
    "doc1 = \"\"\"when we started watching this series on \n",
    "cable i had no idea how addictive it would be \n",
    "even when you hate a character you hold back because \n",
    "they are so beautifully developed you can almost\n",
    "understand why they react to frustration fear greed \n",
    "or temptation the way they do it s almost as if the\n",
    "viewer is experiencing one of christopher s learning \n",
    "curves i can t understand why adriana would put up with \n",
    "christopher s abuse of her verbally physically and \n",
    "emotionally but i just have to read the newspaper to \n",
    "see how many women can and do tolerate such behavior \n",
    "carmella has a dream house endless supply of expensive \n",
    "things but i m sure she would give it up for a loving \n",
    "and faithful husband or maybe not that s why i watch \n",
    "it doesn t matter how many times you watch an episode\n",
    "you can find something you missed the first five times \n",
    "we even watch episodes out of sequence watch season 1 \n",
    "on late night with commercials but all the language a e \n",
    "with language censored reruns on the movie network whenever \n",
    "they re on we re there we ve been totally spoiled now i also \n",
    "love the malaprop s an albacore around my neck is my favorite of\n",
    "johnny boy when these jewels have entered our family vocabulary \n",
    "it is a sign that i should get a life i will when the series\n",
    "ends and i have collected all the dvd s and put the collection \n",
    "in my will\"\"\"\n",
    "doc1 = preprocess(doc1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.7982])"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.predict_proba(np.array([doc1]))[:, 1]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Hashing Vectorizer\n",
    "\n",
    "\n",
    "\n",
    "Convert a collection of text documents to a matrix of deterministic hash token (murmur3) occurrences\n",
    "\n",
    "It turns a collection of text documents into a scipy.sparse matrix holding token occurrence counts (or binary occurrence information), possibly normalized as token frequencies if norm=’l1’ or projected on the euclidean unit sphere if norm=’l2’.\n",
    "\n",
    "Advantages \n",
    "- it is very low memory scalable to large datasets as there is no need to store a vocabulary dictionary in memory\n",
    "- it is fast to pickle and un-pickle as it holds no state besides the constructor parameters\n",
    "- it can be used in a streaming (partial fit) or parallel pipeline as there is no state computed during fit.\n",
    "\n",
    "Disadvantages\n",
    "- there is no way to compute the inverse transform (from feature indices to string feature names) which can be a problem when trying to introspect which features are most important to a model.\n",
    "- there can be collisions: distinct tokens can be mapped to the same feature index. However in practice this is rarely an issue if n_features is large enough (e.g. 2 ** 18 for text classification problems).\n",
    "- no IDF weighting as this would render the transformer stateful."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.5774</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.5774</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>-0.5774</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.3780</td>\n",
       "      <td>0.7559</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>-0.3780</td>\n",
       "      <td>-0.3780</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>-0.5774</td>\n",
       "      <td>-0.5774</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.5774</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       0      1      2      3       4       5       6      7\n",
       "0 0.0000 0.5774 0.0000 0.5774  0.0000  0.0000 -0.5774 0.0000\n",
       "1 0.0000 0.3780 0.7559 0.0000  0.0000 -0.3780 -0.3780 0.0000\n",
       "2 0.0000 0.0000 0.0000 0.0000 -0.5774 -0.5774  0.0000 0.5774"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "hashing_vectorizer = feature_extraction.text.HashingVectorizer(n_features=2 ** 3\n",
    "                                                  , tokenizer=my_tokenizer, ngram_range=(1, 2))\n",
    "\n",
    "corpus = [\"Today is Wednesday\"\n",
    "          , \"Delhi weather is hot today.\"\n",
    "          , \"Delhi roads are not busy in the morning\"]\n",
    "\n",
    "doc_term_matrix = hashing_vectorizer.fit_transform(corpus)\n",
    "\n",
    "pd.DataFrame(doc_term_matrix.toarray()) # Each cell is normalized (l2) row-wise\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 2min 11s, sys: 347 ms, total: 2min 11s\n",
      "Wall time: 2min 2s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "n_features = 2 ** 20\n",
    "\n",
    "hashing_vectorizer = feature_extraction.text.HashingVectorizer(n_features=n_features\n",
    "                                                  , tokenizer=my_tokenizer, ngram_range=(1, 2))\n",
    "X_train_hash = hashing_vectorizer.fit_transform(X_train)\n",
    "X_test_hash = hashing_vectorizer.transform(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((25000, 1048576), (25000, 1048576))"
      ]
     },
     "execution_count": 70,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train_hash.shape, X_test_hash.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training accuracy:  0.91876\n",
      "Test accuracy:  0.87336\n",
      "CPU times: user 18.8 s, sys: 212 ms, total: 19 s\n",
      "Wall time: 3.25 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "lr = linear_model.LogisticRegression(C = 1.0, random_state = 1\n",
    "                            , solver=\"liblinear\")\n",
    "lr.fit(X_train_hash, y_train)\n",
    "y_train_pred = lr.predict(X_train_hash)\n",
    "y_test_pred = lr.predict(X_test_hash)\n",
    "print(\"Training accuracy: \", metrics.accuracy_score(y_train, y_train_pred))\n",
    "print(\"Test accuracy: \", metrics.accuracy_score(y_test, y_test_pred))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Count Vectorizer"
   ]
  },
  {
   "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": "3.6.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}