oyez_controversial.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "f8993439",
   "metadata": {},
   "source": [
    "# Filter Dataset from oyez_gather script\n",
    "This script takes in input oyez.json as generated by the oyez_gather script.\n",
    "Here's an already generated dataset:\n",
    "\n",
    "Minimal JSON compact form (216MB):\n",
    "https://www.dropbox.com/s/9kyk0dr2gf3ls23/oyez.json?dl=0\n",
    "\n",
    "Prettified JSON human-readable form (431 MB):\n",
    "https://www.dropbox.com/s/52a58aac8iujupv/oyez_pretty.json?dl=0"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0ecd86e2",
   "metadata": {},
   "source": [
    "## Imports"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "c4d84fc5",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[nltk_data] Downloading package wordnet to\n",
      "[nltk_data]     C:\\Users\\smitp\\AppData\\Roaming\\nltk_data...\n",
      "[nltk_data]   Package wordnet is already up-to-date!\n",
      "C:\\Users\\smitp\\miniconda3\\lib\\site-packages\\gensim\\similarities\\__init__.py:15: UserWarning: The gensim.similarities.levenshtein submodule is disabled, because the optional Levenshtein package <https://pypi.org/project/python-Levenshtein/> is unavailable. Install Levenhstein (e.g. `pip install python-Levenshtein`) to suppress this warning.\n",
      "  warnings.warn(msg)\n"
     ]
    }
   ],
   "source": [
    "# !pip install json\n",
    "# !pip install pandas\n",
    "# !pip install numpy\n",
    "# !pip install nltk\n",
    "# !pip install bs4\n",
    "# !pip install re\n",
    "# !pip install contractions\n",
    "\n",
    "import json\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import nltk\n",
    "nltk.download('wordnet')\n",
    "import re\n",
    "from bs4 import BeautifulSoup\n",
    "import gensim\n",
    "import contractions\n",
    "from sklearn.svm import LinearSVC\n",
    "from sklearn.metrics import precision_score, recall_score, f1_score, accuracy_score\n",
    "import torch\n",
    "\n",
    "# Just for visuals\n",
    "pd.set_option('display.max_colwidth', None)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "17342c68",
   "metadata": {},
   "source": [
    "# 1. Dataset Preparation"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "40ff3a44",
   "metadata": {},
   "source": [
    "## Load the dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "7cc15010",
   "metadata": {},
   "outputs": [],
   "source": [
    "with open('oyez.json', 'r') as f:\n",
    "    data = json.load(f)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c5ed3f1a",
   "metadata": {},
   "source": [
    "## Filter the dataset to only the columns we need with cases with non-missing values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "c2c9e55a",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Returns true if the given case's judgment has been decided, false otherwise\n",
    "def is_decided(entry):\n",
    "    for timeline in entry['timeline']:\n",
    "        if timeline['event'] == \"Decided\":\n",
    "            return True\n",
    "    return False\n",
    "\n",
    "# Returns true if the given case has the necessary columns for our analysis, false otherwise\n",
    "# Currently only considers:\n",
    "#    1- case's judgment has been decided\n",
    "#    2- case has a non-empty facts field\n",
    "#    3- case has a non-empty decisions field\n",
    "def is_entry_complete(entry):\n",
    "    try:\n",
    "        if not is_decided(entry):\n",
    "            return False\n",
    "        facts = entry['facts_of_the_case']\n",
    "        if facts == None or len(str(facts)) == 0:\n",
    "            return False\n",
    "        if entry['decisions'] == None or len(entry['decisions']) == 0:\n",
    "            return False\n",
    "        return True\n",
    "    except:\n",
    "        return False\n",
    "\n",
    "# Returns a dict with only the necessary columns we need from a given case\n",
    "# Currently the followings columns are considered:\n",
    "#    1-  case ID (assigned by oyez.org)\n",
    "#    2-  case name\n",
    "#    3-  href URL to the oyez.org case\n",
    "#    4-  term (read: year)\n",
    "#    5-  name of the first party\n",
    "#    6-  name of the second party\n",
    "#    7-  facts of the case\n",
    "#    8-  majority vote count\n",
    "#    9-  minority vote count\n",
    "#    10- name of the winning party\n",
    "#    11- the decision type (see below for more detials)\n",
    "#    12- the disposition type (see below for more details)\n",
    "def filter_entry(entry):\n",
    "    row = {}\n",
    "    row['ID'] = entry['ID']\n",
    "    row['name'] = entry['name']\n",
    "    row['href'] = entry['href']\n",
    "    row['term'] = entry['term']\n",
    "    row['first_party'] = entry['first_party']\n",
    "    row['second_party'] = entry['second_party']\n",
    "    row['facts'] = entry['facts_of_the_case']\n",
    "    \n",
    "    # As a simplification, only consider the first decision of the case\n",
    "    # Its rare, but there are cases with multiple decisions.\n",
    "    # How should we handle those? At the moment, we ignore subsequent decisions\n",
    "    decision = entry['decisions'][0]\n",
    "    row['majority_vote'] = decision['majority_vote']\n",
    "    row['minority_vote'] = decision['minority_vote']\n",
    "    row['winning_party'] = decision['winning_party']\n",
    "    row['decision_type'] = decision['decision_type']\n",
    "    row['disposition'] = decision['disposition']\n",
    "    return row\n",
    "\n",
    "# Decision Type can be one the following:\n",
    "#     majority - an opinion in a case that is shared by more than half of the members of a court \n",
    "#     per curiam -  said of a judicial opinion presented as that of the entire court rather than that of any one judge\n",
    "#     plurality opinion - an opinion that received the most votes of any opinion but not enough to be the majority opinion\n",
    "#     equally divided - ?\n",
    "#     dismissal - moot - ?\n",
    "#     dismissal - rule 46 - ?\n",
    "#     dismissal - other - ?\n",
    "#     dismissal - improvidently granted - ?\n",
    "#     memorandum - ?\n",
    "\n",
    "# Disposition can be a combination of the following:\n",
    "#     affirmed - lower court judgment was correct\n",
    "#     reversed - lower court judgment was incorrect\n",
    "#     remanded - send case back to lower court\n",
    "#     vacated - lower court judgment has been cancelled or rendered void"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "e5cd31bd",
   "metadata": {},
   "outputs": [],
   "source": [
    "filtered_data = []\n",
    "for entry in data:\n",
    "    if is_entry_complete(entry):\n",
    "        filtered_data.append(filter_entry(entry))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d815ec89",
   "metadata": {},
   "source": [
    "## Prepare Pandas DataFrame for the Dataset "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "74897779",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "There are 3303 cases.\n"
     ]
    }
   ],
   "source": [
    "df = pd.DataFrame(filtered_data)\n",
    "dfCopy = pd.DataFrame(filtered_data)\n",
    "\n",
    "# Fill empty/None values with an empty string\n",
    "df['winning_party'] = df['winning_party'].fillna('')\n",
    "df['disposition'] = df['disposition'].fillna('')\n",
    "\n",
    "# Remove rows with missing values in the listed columns\n",
    "# df.dropna(inplace=True, subset=['first_party', 'second_party', 'facts'])\n",
    "\n",
    "print(f'There are {len(df)} cases.')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "3059c8c0",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>ID</th>\n",
       "      <th>name</th>\n",
       "      <th>href</th>\n",
       "      <th>term</th>\n",
       "      <th>first_party</th>\n",
       "      <th>second_party</th>\n",
       "      <th>facts</th>\n",
       "      <th>majority_vote</th>\n",
       "      <th>minority_vote</th>\n",
       "      <th>winning_party</th>\n",
       "      <th>decision_type</th>\n",
       "      <th>disposition</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>50606</td>\n",
       "      <td>Roe v. Wade</td>\n",
       "      <td>https://api.oyez.org/cases/1971/70-18</td>\n",
       "      <td>1971</td>\n",
       "      <td>Jane Roe</td>\n",
       "      <td>Henry Wade</td>\n",
       "      <td>&lt;p&gt;In 1970, Jane Roe (a fictional name used in court documents to protect the plaintiff’s identity) filed a lawsuit against Henry Wade, the district attorney of Dallas County, Texas, where she resided, challenging a Texas law making abortion illegal except by a doctor’s orders to save a woman’s life. In her lawsuit, Roe alleged that the state laws were unconstitutionally vague and abridged her right of personal privacy, protected by the First, Fourth, Fifth, Ninth, and Fourteenth Amendments.&lt;/p&gt;\\n</td>\n",
       "      <td>7</td>\n",
       "      <td>2</td>\n",
       "      <td>Jane Roe</td>\n",
       "      <td>majority opinion</td>\n",
       "      <td>reversed</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>50613</td>\n",
       "      <td>Stanley v. Illinois</td>\n",
       "      <td>https://api.oyez.org/cases/1971/70-5014</td>\n",
       "      <td>1971</td>\n",
       "      <td>Peter Stanley, Sr.</td>\n",
       "      <td>Illinois</td>\n",
       "      <td>&lt;p&gt;Joan Stanley had three children with Peter Stanley.  The Stanleys never married, but lived together off and on for 18 years.  When Joan died, the State of Illinois took the children.  Under Illinois law, unwed fathers were presumed unfit parents regardless of their actual fitness and their children became wards of the state.  Peter appealed the decision, arguing that the Illinois law violated the Equal Protection Clause of the Fourteenth Amendment because unwed mothers were not deprived of their children without a showing that they were actually unfit parents.  The Illinois Supreme Court rejected Stanley’s Equal Protection claim, holding that his actual fitness as a parent was irrelevant because he and the children’s mother were unmarried.&lt;/p&gt;\\n</td>\n",
       "      <td>5</td>\n",
       "      <td>2</td>\n",
       "      <td>Stanley</td>\n",
       "      <td>majority opinion</td>\n",
       "      <td>reversed/remanded</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>50623</td>\n",
       "      <td>Giglio v. United States</td>\n",
       "      <td>https://api.oyez.org/cases/1971/70-29</td>\n",
       "      <td>1971</td>\n",
       "      <td>John Giglio</td>\n",
       "      <td>United States</td>\n",
       "      <td>&lt;p&gt;John Giglio was convicted of passing forged money orders.  While his appeal to the U.S. Court of Appeals for the Second Circuit was pending, Giglio’s counsel discovered new evidence. The evidence indicated that the prosecution failed to disclose that it promised a key witness immunity from prosecution in exchange for testimony against Giglio.  The district court denied Giglio’s motion for a new trial, finding that the error did not affect the verdict.  The Court of Appeals affirmed.&lt;/p&gt;\\n</td>\n",
       "      <td>7</td>\n",
       "      <td>0</td>\n",
       "      <td>Giglio</td>\n",
       "      <td>majority opinion</td>\n",
       "      <td>reversed/remanded</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      ID                     name                                     href  \\\n",
       "0  50606              Roe v. Wade    https://api.oyez.org/cases/1971/70-18   \n",
       "1  50613      Stanley v. Illinois  https://api.oyez.org/cases/1971/70-5014   \n",
       "2  50623  Giglio v. United States    https://api.oyez.org/cases/1971/70-29   \n",
       "\n",
       "   term          first_party   second_party  \\\n",
       "0  1971             Jane Roe     Henry Wade   \n",
       "1  1971  Peter Stanley, Sr.        Illinois   \n",
       "2  1971         John Giglio   United States   \n",
       "\n",
       "                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    facts  \\\n",
       "0                                                                                                                                                                                                                                                                  <p>In 1970, Jane Roe (a fictional name used in court documents to protect the plaintiff’s identity) filed a lawsuit against Henry Wade, the district attorney of Dallas County, Texas, where she resided, challenging a Texas law making abortion illegal except by a doctor’s orders to save a woman’s life. In her lawsuit, Roe alleged that the state laws were unconstitutionally vague and abridged her right of personal privacy, protected by the First, Fourth, Fifth, Ninth, and Fourteenth Amendments.</p>\\n   \n",
       "1  <p>Joan Stanley had three children with Peter Stanley.  The Stanleys never married, but lived together off and on for 18 years.  When Joan died, the State of Illinois took the children.  Under Illinois law, unwed fathers were presumed unfit parents regardless of their actual fitness and their children became wards of the state.  Peter appealed the decision, arguing that the Illinois law violated the Equal Protection Clause of the Fourteenth Amendment because unwed mothers were not deprived of their children without a showing that they were actually unfit parents.  The Illinois Supreme Court rejected Stanley’s Equal Protection claim, holding that his actual fitness as a parent was irrelevant because he and the children’s mother were unmarried.</p>\\n   \n",
       "2                                                                                                                                                                                                                                                                        <p>John Giglio was convicted of passing forged money orders.  While his appeal to the U.S. Court of Appeals for the Second Circuit was pending, Giglio’s counsel discovered new evidence. The evidence indicated that the prosecution failed to disclose that it promised a key witness immunity from prosecution in exchange for testimony against Giglio.  The district court denied Giglio’s motion for a new trial, finding that the error did not affect the verdict.  The Court of Appeals affirmed.</p>\\n   \n",
       "\n",
       "   majority_vote  minority_vote winning_party     decision_type  \\\n",
       "0              7              2      Jane Roe  majority opinion   \n",
       "1              5              2       Stanley  majority opinion   \n",
       "2              7              0        Giglio  majority opinion   \n",
       "\n",
       "         disposition  \n",
       "0           reversed  \n",
       "1  reversed/remanded  \n",
       "2  reversed/remanded  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "display(df.head(n=3))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f3f7f3f5",
   "metadata": {},
   "source": [
    "# 2. Preprocess Dataset"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fbeec309",
   "metadata": {},
   "source": [
    "## Statistics before Preprocessing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "cbc0fcd9",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Average facts character length (before preprocesesing): 1112\n"
     ]
    }
   ],
   "source": [
    "avg_before_preprocessing = df['facts'].apply(lambda x: len(str(x))).mean()\n",
    "print(f'Average facts character length (before preprocesesing): {avg_before_preprocessing:.0f}')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1af7fe6e",
   "metadata": {},
   "source": [
    "## Lowercase"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "d574deb5",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "172.37874659400546\n"
     ]
    }
   ],
   "source": [
    "df['facts'] = df['facts'].str.lower()\n",
    "dfCopy['facts'] = dfCopy['facts'].str.lower()\n",
    "\n",
    "print(np.average(dfCopy['facts'].apply(lambda x: len(x.split()))))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "54303c9c",
   "metadata": {},
   "source": [
    "## Remove HTML/URL"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "e3945388",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Function to remove HTML tags and URLs from a string\n",
    "def sanitize_review(text):\n",
    "    # remove HTML tags\n",
    "    text = BeautifulSoup(str(text), 'html.parser').get_text()   \n",
    "    # remove URLS\n",
    "    text = re.sub(r'http\\S+', '', str(text))\n",
    "    return text\n",
    "\n",
    "df['facts'] = df['facts'].apply(sanitize_review)\n",
    "dfCopy['facts'] = dfCopy['facts'].apply(sanitize_review)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "74009061",
   "metadata": {},
   "source": [
    "## Remove Contractions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "4d8fd034",
   "metadata": {},
   "outputs": [],
   "source": [
    "def fix_contractions(text):\n",
    "    return contractions.fix(text)\n",
    "\n",
    "df['facts'] = df['facts'].apply(fix_contractions)\n",
    "dfCopy['facts'] = dfCopy['facts'].apply(fix_contractions)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1d4297c3",
   "metadata": {},
   "source": [
    "## Remove non-alphabetical characters"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "82896ea8",
   "metadata": {},
   "outputs": [],
   "source": [
    "df['facts'] = df['facts'].str.replace('[^a-zA-Z\\s]', ' ')\n",
    "dfCopy['facts'] = dfCopy['facts'].str.replace('[^a-zA-Z\\s]', ' ')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "02c2d53c",
   "metadata": {},
   "source": [
    "## Remove extra spaces"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "bcf32717",
   "metadata": {},
   "outputs": [],
   "source": [
    "def remove_extra_spaces(text):\n",
    "    return ' '.join(str(text).split())\n",
    "\n",
    "df['facts'] = df['facts'].apply(remove_extra_spaces)\n",
    "dfCopy['facts'] = dfCopy['facts'].apply(remove_extra_spaces)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "080d0fd6",
   "metadata": {},
   "source": [
    "## Remove stop words"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "38548a1b",
   "metadata": {},
   "outputs": [],
   "source": [
    "from nltk.corpus import stopwords\n",
    "nltk.download('stopwords')\n",
    "stop = stopwords.words('english')\n",
    "\n",
    "def remove_stop_words(text):\n",
    "    return ' '.join([word for word in str(text).split() if word not in (stop)])\n",
    "\n",
    "df['facts'] = df['facts'].apply(remove_stop_words)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1498fc6a",
   "metadata": {},
   "source": [
    "## Lemmatization"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "fc1d20e4",
   "metadata": {},
   "outputs": [],
   "source": [
    "from nltk.stem import WordNetLemmatizer\n",
    "\n",
    "tokenizer = nltk.tokenize.WhitespaceTokenizer()\n",
    "lemmatizer = WordNetLemmatizer()\n",
    "\n",
    "def lemmatize(text):\n",
    "    return ' '.join([lemmatizer.lemmatize(word) for word in tokenizer.tokenize(text)])\n",
    "\n",
    "df['facts'] = df['facts'].apply(lemmatize)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "97a21e92",
   "metadata": {},
   "source": [
    "## Statistics after Preprocessing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "761f95bd",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Average facts character length (after preprocesesing): 776\n"
     ]
    }
   ],
   "source": [
    "avg_after_preprocessing = df['facts'].apply(lambda x: len(str(x))).mean()\n",
    "print(f'Average facts character length (after preprocesesing): {avg_after_preprocessing:.0f}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "b2efc36a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>ID</th>\n",
       "      <th>name</th>\n",
       "      <th>href</th>\n",
       "      <th>term</th>\n",
       "      <th>first_party</th>\n",
       "      <th>second_party</th>\n",
       "      <th>facts</th>\n",
       "      <th>majority_vote</th>\n",
       "      <th>minority_vote</th>\n",
       "      <th>winning_party</th>\n",
       "      <th>decision_type</th>\n",
       "      <th>disposition</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>50606</td>\n",
       "      <td>Roe v. Wade</td>\n",
       "      <td>https://api.oyez.org/cases/1971/70-18</td>\n",
       "      <td>1971</td>\n",
       "      <td>Jane Roe</td>\n",
       "      <td>Henry Wade</td>\n",
       "      <td>jane roe fictional name used court document protect plaintiff identity filed lawsuit henry wade district attorney dallas county texas resided challenging texas law making abortion illegal except doctor order save woman life lawsuit roe alleged state law unconstitutionally vague abridged right personal privacy protected first fourth fifth ninth fourteenth amendment</td>\n",
       "      <td>7</td>\n",
       "      <td>2</td>\n",
       "      <td>Jane Roe</td>\n",
       "      <td>majority opinion</td>\n",
       "      <td>reversed</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>50613</td>\n",
       "      <td>Stanley v. Illinois</td>\n",
       "      <td>https://api.oyez.org/cases/1971/70-5014</td>\n",
       "      <td>1971</td>\n",
       "      <td>Peter Stanley, Sr.</td>\n",
       "      <td>Illinois</td>\n",
       "      <td>joan stanley three child peter stanley stanley never married lived together year joan died state illinois took child illinois law unwed father presumed unfit parent regardless actual fitness child became ward state peter appealed decision arguing illinois law violated equal protection clause fourteenth amendment unwed mother deprived child without showing actually unfit parent illinois supreme court rejected stanley equal protection claim holding actual fitness parent irrelevant child mother unmarried</td>\n",
       "      <td>5</td>\n",
       "      <td>2</td>\n",
       "      <td>Stanley</td>\n",
       "      <td>majority opinion</td>\n",
       "      <td>reversed/remanded</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>50623</td>\n",
       "      <td>Giglio v. United States</td>\n",
       "      <td>https://api.oyez.org/cases/1971/70-29</td>\n",
       "      <td>1971</td>\n",
       "      <td>John Giglio</td>\n",
       "      <td>United States</td>\n",
       "      <td>john giglio convicted passing forged money order appeal court appeal second circuit pending giglio counsel discovered new evidence evidence indicated prosecution failed disclose promised key witness immunity prosecution exchange testimony giglio district court denied giglio motion new trial finding error affect verdict court appeal affirmed</td>\n",
       "      <td>7</td>\n",
       "      <td>0</td>\n",
       "      <td>Giglio</td>\n",
       "      <td>majority opinion</td>\n",
       "      <td>reversed/remanded</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      ID                     name                                     href  \\\n",
       "0  50606              Roe v. Wade    https://api.oyez.org/cases/1971/70-18   \n",
       "1  50613      Stanley v. Illinois  https://api.oyez.org/cases/1971/70-5014   \n",
       "2  50623  Giglio v. United States    https://api.oyez.org/cases/1971/70-29   \n",
       "\n",
       "   term          first_party   second_party  \\\n",
       "0  1971             Jane Roe     Henry Wade   \n",
       "1  1971  Peter Stanley, Sr.        Illinois   \n",
       "2  1971         John Giglio   United States   \n",
       "\n",
       "                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        facts  \\\n",
       "0                                                                                                                                              jane roe fictional name used court document protect plaintiff identity filed lawsuit henry wade district attorney dallas county texas resided challenging texas law making abortion illegal except doctor order save woman life lawsuit roe alleged state law unconstitutionally vague abridged right personal privacy protected first fourth fifth ninth fourteenth amendment   \n",
       "1  joan stanley three child peter stanley stanley never married lived together year joan died state illinois took child illinois law unwed father presumed unfit parent regardless actual fitness child became ward state peter appealed decision arguing illinois law violated equal protection clause fourteenth amendment unwed mother deprived child without showing actually unfit parent illinois supreme court rejected stanley equal protection claim holding actual fitness parent irrelevant child mother unmarried   \n",
       "2                                                                                                                                                                      john giglio convicted passing forged money order appeal court appeal second circuit pending giglio counsel discovered new evidence evidence indicated prosecution failed disclose promised key witness immunity prosecution exchange testimony giglio district court denied giglio motion new trial finding error affect verdict court appeal affirmed   \n",
       "\n",
       "   majority_vote  minority_vote winning_party     decision_type  \\\n",
       "0              7              2      Jane Roe  majority opinion   \n",
       "1              5              2       Stanley  majority opinion   \n",
       "2              7              0        Giglio  majority opinion   \n",
       "\n",
       "         disposition  \n",
       "0           reversed  \n",
       "1  reversed/remanded  \n",
       "2  reversed/remanded  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "display(df.head(n=3))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "41f4afca",
   "metadata": {},
   "source": [
    "# 3. How controversial is a case?\n",
    "Based on the votes distribution and judge opinions, determine how controversial or disputed a case is."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "1dee78b8",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Extract only the columns we need\n",
    "df2 = df[['ID', 'href', 'name', 'facts', 'majority_vote', 'minority_vote']].copy()\n",
    "\n",
    "# Remove rows with no majority_vote or no minority_vote\n",
    "df2.dropna(subset=['majority_vote', 'minority_vote'], inplace=True)\n",
    "\n",
    "# Given a row, return the controversial class, or -1 if its an invalid row (i.e. 0, 0 for majority/minority votes)\n",
    "'''\n",
    "    Returns:\n",
    "        -1 - invalid row\n",
    "         0 - not controversial\n",
    "         1 - controversial\n",
    "'''\n",
    "def label_row(row):\n",
    "    majority = row['majority_vote']\n",
    "    minority = row['minority_vote']\n",
    "    total = majority + minority\n",
    "    if total == 0:\n",
    "        return -1\n",
    "    else:\n",
    "        controversial = 1 - majority / total\n",
    "        # The threshold is 0.3 rather than 0.5\n",
    "        # You may be tempted to set this to 0.5, but only ~16 cases have controversial value of [0.5, 1.0]\n",
    "        if controversial <= 0.3:\n",
    "            return 0\n",
    "        else:\n",
    "            return 1\n",
    "\n",
    "# Create a new column for the controversial class for each row\n",
    "df2['controversial'] = df2.apply(lambda row: label_row(row), axis=1)\n",
    "\n",
    "# Remove rows with invalid controversial class\n",
    "df2 = df2[df2['controversial'] != -1]\n",
    "\n",
    "# Remove unneeded columns\n",
    "df2.drop(['majority_vote', 'minority_vote'], axis=1, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "90d3a405",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>ID</th>\n",
       "      <th>href</th>\n",
       "      <th>name</th>\n",
       "      <th>facts</th>\n",
       "      <th>controversial</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>50606</td>\n",
       "      <td>https://api.oyez.org/cases/1971/70-18</td>\n",
       "      <td>Roe v. Wade</td>\n",
       "      <td>jane roe fictional name used court document protect plaintiff identity filed lawsuit henry wade district attorney dallas county texas resided challenging texas law making abortion illegal except doctor order save woman life lawsuit roe alleged state law unconstitutionally vague abridged right personal privacy protected first fourth fifth ninth fourteenth amendment</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>50613</td>\n",
       "      <td>https://api.oyez.org/cases/1971/70-5014</td>\n",
       "      <td>Stanley v. Illinois</td>\n",
       "      <td>joan stanley three child peter stanley stanley never married lived together year joan died state illinois took child illinois law unwed father presumed unfit parent regardless actual fitness child became ward state peter appealed decision arguing illinois law violated equal protection clause fourteenth amendment unwed mother deprived child without showing actually unfit parent illinois supreme court rejected stanley equal protection claim holding actual fitness parent irrelevant child mother unmarried</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>50623</td>\n",
       "      <td>https://api.oyez.org/cases/1971/70-29</td>\n",
       "      <td>Giglio v. United States</td>\n",
       "      <td>john giglio convicted passing forged money order appeal court appeal second circuit pending giglio counsel discovered new evidence evidence indicated prosecution failed disclose promised key witness immunity prosecution exchange testimony giglio district court denied giglio motion new trial finding error affect verdict court appeal affirmed</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      ID                                     href                     name  \\\n",
       "0  50606    https://api.oyez.org/cases/1971/70-18              Roe v. Wade   \n",
       "1  50613  https://api.oyez.org/cases/1971/70-5014      Stanley v. Illinois   \n",
       "2  50623    https://api.oyez.org/cases/1971/70-29  Giglio v. United States   \n",
       "\n",
       "                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        facts  \\\n",
       "0                                                                                                                                              jane roe fictional name used court document protect plaintiff identity filed lawsuit henry wade district attorney dallas county texas resided challenging texas law making abortion illegal except doctor order save woman life lawsuit roe alleged state law unconstitutionally vague abridged right personal privacy protected first fourth fifth ninth fourteenth amendment   \n",
       "1  joan stanley three child peter stanley stanley never married lived together year joan died state illinois took child illinois law unwed father presumed unfit parent regardless actual fitness child became ward state peter appealed decision arguing illinois law violated equal protection clause fourteenth amendment unwed mother deprived child without showing actually unfit parent illinois supreme court rejected stanley equal protection claim holding actual fitness parent irrelevant child mother unmarried   \n",
       "2                                                                                                                                                                      john giglio convicted passing forged money order appeal court appeal second circuit pending giglio counsel discovered new evidence evidence indicated prosecution failed disclose promised key witness immunity prosecution exchange testimony giglio district court denied giglio motion new trial finding error affect verdict court appeal affirmed   \n",
       "\n",
       "   controversial  \n",
       "0              0  \n",
       "1              0  \n",
       "2              0  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "display(df2.head(n=3))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "619ffd7d",
   "metadata": {},
   "source": [
    "## Display frequency of each controversial class"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "298b7ee0",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0 1]\n",
      "[1976 1241]\n"
     ]
    }
   ],
   "source": [
    "index, counts = np.unique(df2['controversial'].values, return_counts=True)\n",
    "print(index) # classes\n",
    "print(counts) # frequency"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "d7490b77-5628-489a-b48d-5b443b1227a8",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>ID</th>\n",
       "      <th>name</th>\n",
       "      <th>href</th>\n",
       "      <th>term</th>\n",
       "      <th>first_party</th>\n",
       "      <th>second_party</th>\n",
       "      <th>facts</th>\n",
       "      <th>majority_vote</th>\n",
       "      <th>minority_vote</th>\n",
       "      <th>winning_party</th>\n",
       "      <th>decision_type</th>\n",
       "      <th>disposition</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>50606</td>\n",
       "      <td>Roe v. Wade</td>\n",
       "      <td>https://api.oyez.org/cases/1971/70-18</td>\n",
       "      <td>1971</td>\n",
       "      <td>Jane Roe</td>\n",
       "      <td>Henry Wade</td>\n",
       "      <td>in jane roe a fictional name used in court documents to protect the plaintiff s identity filed a lawsuit against henry wade the district attorney of dallas county texas where she resided challenging a texas law making abortion illegal except by a doctor s orders to save a woman s life in her lawsuit roe alleged that the state laws were unconstitutionally vague and abridged her right of personal privacy protected by the first fourth fifth ninth and fourteenth amendments</td>\n",
       "      <td>7</td>\n",
       "      <td>2</td>\n",
       "      <td>Jane Roe</td>\n",
       "      <td>majority opinion</td>\n",
       "      <td>reversed</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>50613</td>\n",
       "      <td>Stanley v. Illinois</td>\n",
       "      <td>https://api.oyez.org/cases/1971/70-5014</td>\n",
       "      <td>1971</td>\n",
       "      <td>Peter Stanley, Sr.</td>\n",
       "      <td>Illinois</td>\n",
       "      <td>joan stanley had three children with peter stanley the stanleys never married but lived together off and on for years when joan died the state of illinois took the children under illinois law unwed fathers were presumed unfit parents regardless of their actual fitness and their children became wards of the state peter appealed the decision arguing that the illinois law violated the equal protection clause of the fourteenth amendment because unwed mothers were not deprived of their children without a showing that they were actually unfit parents the illinois supreme court rejected stanley s equal protection claim holding that his actual fitness as a parent was irrelevant because he and the children s mother were unmarried</td>\n",
       "      <td>5</td>\n",
       "      <td>2</td>\n",
       "      <td>Stanley</td>\n",
       "      <td>majority opinion</td>\n",
       "      <td>reversed/remanded</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>50623</td>\n",
       "      <td>Giglio v. United States</td>\n",
       "      <td>https://api.oyez.org/cases/1971/70-29</td>\n",
       "      <td>1971</td>\n",
       "      <td>John Giglio</td>\n",
       "      <td>United States</td>\n",
       "      <td>john giglio was convicted of passing forged money orders while his appeal to the you s court of appeals for the second circuit was pending giglio s counsel discovered new evidence the evidence indicated that the prosecution failed to disclose that it promised a key witness immunity from prosecution in exchange for testimony against giglio the district court denied giglio s motion for a new trial finding that the error did not affect the verdict the court of appeals affirmed</td>\n",
       "      <td>7</td>\n",
       "      <td>0</td>\n",
       "      <td>Giglio</td>\n",
       "      <td>majority opinion</td>\n",
       "      <td>reversed/remanded</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      ID                     name                                     href  \\\n",
       "0  50606              Roe v. Wade    https://api.oyez.org/cases/1971/70-18   \n",
       "1  50613      Stanley v. Illinois  https://api.oyez.org/cases/1971/70-5014   \n",
       "2  50623  Giglio v. United States    https://api.oyez.org/cases/1971/70-29   \n",
       "\n",
       "   term          first_party   second_party  \\\n",
       "0  1971             Jane Roe     Henry Wade   \n",
       "1  1971  Peter Stanley, Sr.        Illinois   \n",
       "2  1971         John Giglio   United States   \n",
       "\n",
       "                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        facts  \\\n",
       "0                                                                                                                                                                                                                                                                   in jane roe a fictional name used in court documents to protect the plaintiff s identity filed a lawsuit against henry wade the district attorney of dallas county texas where she resided challenging a texas law making abortion illegal except by a doctor s orders to save a woman s life in her lawsuit roe alleged that the state laws were unconstitutionally vague and abridged her right of personal privacy protected by the first fourth fifth ninth and fourteenth amendments   \n",
       "1  joan stanley had three children with peter stanley the stanleys never married but lived together off and on for years when joan died the state of illinois took the children under illinois law unwed fathers were presumed unfit parents regardless of their actual fitness and their children became wards of the state peter appealed the decision arguing that the illinois law violated the equal protection clause of the fourteenth amendment because unwed mothers were not deprived of their children without a showing that they were actually unfit parents the illinois supreme court rejected stanley s equal protection claim holding that his actual fitness as a parent was irrelevant because he and the children s mother were unmarried   \n",
       "2                                                                                                                                                                                                                                                              john giglio was convicted of passing forged money orders while his appeal to the you s court of appeals for the second circuit was pending giglio s counsel discovered new evidence the evidence indicated that the prosecution failed to disclose that it promised a key witness immunity from prosecution in exchange for testimony against giglio the district court denied giglio s motion for a new trial finding that the error did not affect the verdict the court of appeals affirmed   \n",
       "\n",
       "   majority_vote  minority_vote winning_party     decision_type  \\\n",
       "0              7              2      Jane Roe  majority opinion   \n",
       "1              5              2       Stanley  majority opinion   \n",
       "2              7              0        Giglio  majority opinion   \n",
       "\n",
       "         disposition  \n",
       "0           reversed  \n",
       "1  reversed/remanded  \n",
       "2  reversed/remanded  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "display(dfCopy.head(n=3))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1a188d21-9c62-4b1c-9b70-77af9db62c15",
   "metadata": {},
   "source": [
    "### Train Word2Vec Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "a341485b-4e2c-4518-9678-09992f143037",
   "metadata": {},
   "outputs": [],
   "source": [
    "word2vec = gensim.models.Word2Vec(sentences = dfCopy['facts'].apply(lambda x: x.split()), vector_size=100, window=11)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "98fc16f7-2e15-4d76-964a-59f2cfc79125",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>ID</th>\n",
       "      <th>href</th>\n",
       "      <th>name</th>\n",
       "      <th>facts</th>\n",
       "      <th>controversial</th>\n",
       "      <th>avgWord2Vec</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>50606</td>\n",
       "      <td>https://api.oyez.org/cases/1971/70-18</td>\n",
       "      <td>Roe v. Wade</td>\n",
       "      <td>jane roe fictional name used court document protect plaintiff identity filed lawsuit henry wade district attorney dallas county texas resided challenging texas law making abortion illegal except doctor order save woman life lawsuit roe alleged state law unconstitutionally vague abridged right personal privacy protected first fourth fifth ninth fourteenth amendment</td>\n",
       "      <td>0</td>\n",
       "      <td>[-0.2147746344655752, 0.1555396685997645, 0.37713931924353045, 0.27548077730617176, 0.2836545525351539, -0.002906513847410679, 0.420634981294473, 0.26051431926433, -0.11545287030438582, 0.2326540336990729, 0.11856872833644351, -0.2924590972562631, 0.33246960045769813, 0.5163163010776043, -0.3086835518262039, 0.059263788672784964, -0.05673542956278349, -0.15400740347802638, -0.027797024132062993, -0.6733945586780707, 0.19539391488729355, 0.010653671749557059, 0.2179606451342503, -0.11485163840154806, 0.09168348658209045, 0.00015018497904141745, 0.1419217615785601, -0.11161981110305835, -0.8370522925754388, -0.25504131921799855, 0.06922466423362493, 0.005723467571660876, -0.05825177914307763, -0.3733747008939584, 0.10120032807812095, -0.1505576040269807, -0.0009496871226777633, -0.021931776873146493, -0.22414211375017962, -0.7134683486322562, -0.02965872305445373, -0.2591561682522297, -0.1708799807416896, 0.31113021649420264, 0.18806260467196503, -0.3773106664450218, -0.43701099207003913, 0.161124746829737, 0.19853670397152504, 0.1384008579258807, -0.12626477303138622, 0.1128284882629911, 0.08823897570371628, 0.14573298244737087, -0.15123370630084537, -0.10183177204181751, 0.11699323664108913, -0.2152861072557668, -0.4109185595600866, 0.04488039909861982, 0.29730659667402504, -0.02103337547198559, 0.10366517174988985, -0.10990773168824186, -0.6709972395002842, 0.4728614409516255, -0.2619458982363964, 0.1063814707348744, -0.5414489042262236, 0.019532115377951412, -0.5057210169235865, 0.5179418066889048, 0.19314716504265864, -0.14546642091746131, 0.45177034541964534, -0.11592952281236649, 0.4537292083601157, 0.03459292645178114, -0.4578340260932843, -0.1954996273604532, -0.0006656653185685476, -0.05180199009055893, -0.009347545423855384, 0.5512251868844033, -0.19694734343638023, -0.00985147464205511, 0.2168656200590582, -0.0035145952738821506, 0.2092997016128114, 0.4424186132242903, -0.21791437707996617, 0.13925800899975002, 0.4881681837389866, 0.1291978319020321, 0.6021646431088448, 0.15163587469607592, -0.014843907452498873, 0.10354407933695862, -0.04580927594875296, 0.3084240119997412]</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      ID                                   href         name  \\\n",
       "0  50606  https://api.oyez.org/cases/1971/70-18  Roe v. Wade   \n",
       "\n",
       "                                                                                                                                                                                                                                                                                                                                                                            facts  \\\n",
       "0  jane roe fictional name used court document protect plaintiff identity filed lawsuit henry wade district attorney dallas county texas resided challenging texas law making abortion illegal except doctor order save woman life lawsuit roe alleged state law unconstitutionally vague abridged right personal privacy protected first fourth fifth ninth fourteenth amendment   \n",
       "\n",
       "   controversial  \\\n",
       "0              0   \n",
       "\n",
       "                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 avgWord2Vec  \n",
       "0  [-0.2147746344655752, 0.1555396685997645, 0.37713931924353045, 0.27548077730617176, 0.2836545525351539, -0.002906513847410679, 0.420634981294473, 0.26051431926433, -0.11545287030438582, 0.2326540336990729, 0.11856872833644351, -0.2924590972562631, 0.33246960045769813, 0.5163163010776043, -0.3086835518262039, 0.059263788672784964, -0.05673542956278349, -0.15400740347802638, -0.027797024132062993, -0.6733945586780707, 0.19539391488729355, 0.010653671749557059, 0.2179606451342503, -0.11485163840154806, 0.09168348658209045, 0.00015018497904141745, 0.1419217615785601, -0.11161981110305835, -0.8370522925754388, -0.25504131921799855, 0.06922466423362493, 0.005723467571660876, -0.05825177914307763, -0.3733747008939584, 0.10120032807812095, -0.1505576040269807, -0.0009496871226777633, -0.021931776873146493, -0.22414211375017962, -0.7134683486322562, -0.02965872305445373, -0.2591561682522297, -0.1708799807416896, 0.31113021649420264, 0.18806260467196503, -0.3773106664450218, -0.43701099207003913, 0.161124746829737, 0.19853670397152504, 0.1384008579258807, -0.12626477303138622, 0.1128284882629911, 0.08823897570371628, 0.14573298244737087, -0.15123370630084537, -0.10183177204181751, 0.11699323664108913, -0.2152861072557668, -0.4109185595600866, 0.04488039909861982, 0.29730659667402504, -0.02103337547198559, 0.10366517174988985, -0.10990773168824186, -0.6709972395002842, 0.4728614409516255, -0.2619458982363964, 0.1063814707348744, -0.5414489042262236, 0.019532115377951412, -0.5057210169235865, 0.5179418066889048, 0.19314716504265864, -0.14546642091746131, 0.45177034541964534, -0.11592952281236649, 0.4537292083601157, 0.03459292645178114, -0.4578340260932843, -0.1954996273604532, -0.0006656653185685476, -0.05180199009055893, -0.009347545423855384, 0.5512251868844033, -0.19694734343638023, -0.00985147464205511, 0.2168656200590582, -0.0035145952738821506, 0.2092997016128114, 0.4424186132242903, -0.21791437707996617, 0.13925800899975002, 0.4881681837389866, 0.1291978319020321, 0.6021646431088448, 0.15163587469607592, -0.014843907452498873, 0.10354407933695862, -0.04580927594875296, 0.3084240119997412]  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "def calculateAvgVector(review):\n",
    "    avgVec = np.zeros(100)\n",
    "    num = 0\n",
    "    for word in review:\n",
    "        if word in word2vec.wv:\n",
    "            avgVec += word2vec.wv[word]\n",
    "            num += 1\n",
    "    avgVec /= num\n",
    "    return avgVec\n",
    "\n",
    "\n",
    "df2['avgWord2Vec'] = df['facts'].apply(calculateAvgVector)\n",
    "\n",
    "display(df2.head(n=1))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d62c3a9a-e53d-4d10-bb75-f99eb64a86d2",
   "metadata": {},
   "source": [
    "### Test Basic Linear Classifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "a24e1625-2fbf-447a-abf1-11d9f1859f6f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LinearSVC(random_state=0)"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "svc = LinearSVC(random_state=0)\n",
    "svc.fit(df2['avgWord2Vec'].values.tolist(), df2['controversial'].values.tolist())\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "1fb74818-06cb-4398-892e-a7afecd9d0b6",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.6151694124961143\n",
      "{0: 3214, 1: 3}\n"
     ]
    }
   ],
   "source": [
    "predictions = svc.predict(df2['avgWord2Vec'].values.tolist())\n",
    "counts = {}\n",
    "for item in predictions:\n",
    "    if item not in counts:\n",
    "        counts[item] = 1\n",
    "    else:\n",
    "        counts[item] += 1\n",
    "\n",
    "print(accuracy_score(predictions, df2['controversial'].values.tolist()))\n",
    "print(counts)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "5771ea45-78b1-421a-99a2-c2a28d6918cf",
   "metadata": {},
   "outputs": [],
   "source": [
    "maxFactLen = max(dfCopy['facts'].apply(lambda x: len(x.split(\" \"))))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "5d7ca750-d001-44eb-ae9e-1a5ae65fa579",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "952\n"
     ]
    }
   ],
   "source": [
    "def concatWord2Vec(item):\n",
    "    if len(item.split()) >= maxFactLen:\n",
    "        print(len(item.split(\" \"))) \n",
    "    vecList = []\n",
    "    for word in item.split():\n",
    "        if word in word2vec.wv:\n",
    "            vecList.append(word2vec.wv[word])\n",
    "    while len(vecList) < maxFactLen:\n",
    "        vecList.append([0 for i in range(100)])\n",
    "    if len(vecList) != 952:\n",
    "        print(len(vecList))\n",
    "    return vecList\n",
    "\n",
    "df2['concatWord2Vec'] = dfCopy['facts'].apply(concatWord2Vec)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "6b5c2c7e-e4dc-4dfc-bbd9-17c21e4e6df3",
   "metadata": {},
   "outputs": [],
   "source": [
    "class BLSTM(torch.nn.Module):\n",
    "    def __init__(self):\n",
    "        super(BLSTM, self).__init__()\n",
    "        self.lstm = torch.nn.LSTM(input_size=100,hidden_size=256,num_layers=1,batch_first=True,dropout=0.33, bidirectional=True).cuda()\n",
    "        self.linear = torch.nn.Linear(512, 128).cuda()\n",
    "        self.classify = torch.nn.Linear(128, 2).cuda()\n",
    "        self.elu = torch.nn.ELU()\n",
    "        self.softmax = torch.nn.Softmax(dim=2)\n",
    "        \n",
    "    def forward(self, x):\n",
    "        initHidden = torch.zeros(2, x.size(0), 256).type(torch.cuda.FloatTensor)\n",
    "        initInternal = torch.zeros(2, x.size(0), 256).type(torch.cuda.FloatTensor)\n",
    "        initialOutput, hidden = self.lstm(x.float(), (initHidden, initInternal))\n",
    "        out = self.linear(initialOutput)\n",
    "        out = self.elu(out)\n",
    "        out = self.classify(out)\n",
    "        return out"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "16586490-9ae0-4254-a61c-9b27b6dfa496",
   "metadata": {},
   "outputs": [],
   "source": [
    "finalLabels = [[df2['controversial'].values.tolist()[i] for j in range(952)] for i in range(len(df2['controversial'].values.tolist()))]\n",
    "\n",
    "trainingVectors = torch.as_tensor(df2['concatWord2Vec'].values.tolist()).cuda()\n",
    "trainingLabels = torch.flatten(torch.as_tensor(finalLabels)).cuda()\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "689970b3-ee44-4e9a-bea8-a67f8b1dd023",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    0\n",
       "1    0\n",
       "2    0\n",
       "Name: controversial, dtype: int64"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "display(df2['controversial'].head(n=3))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "2c35d4cb-744c-4378-b7a3-26cdba36e444",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\smitp\\miniconda3\\lib\\site-packages\\torch\\nn\\modules\\rnn.py:62: UserWarning: dropout option adds dropout after all but last recurrent layer, so non-zero dropout expects num_layers greater than 1, but got dropout=0.33 and num_layers=1\n",
      "  warnings.warn(\"dropout option adds dropout after all but last \"\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tensor(0.6889, device='cuda:0', grad_fn=<NllLossBackward>)\n"
     ]
    },
    {
     "ename": "RuntimeError",
     "evalue": "CUDA out of memory. Tried to allocate 1.47 GiB (GPU 0; 8.00 GiB total capacity; 5.26 GiB already allocated; 468.12 MiB free; 5.68 GiB reserved in total by PyTorch)",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mRuntimeError\u001b[0m                              Traceback (most recent call last)",
      "\u001b[1;32m~\\AppData\\Local\\Temp/ipykernel_22296/2233348646.py\u001b[0m in \u001b[0;36m<module>\u001b[1;34m\u001b[0m\n\u001b[0;32m     26\u001b[0m     \u001b[1;32mif\u001b[0m \u001b[0mi\u001b[0m\u001b[1;33m%\u001b[0m\u001b[1;36m10\u001b[0m \u001b[1;33m==\u001b[0m \u001b[1;36m0\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m     27\u001b[0m         \u001b[0mprint\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mloss\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m---> 28\u001b[1;33m     \u001b[0mloss\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mbackward\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m     29\u001b[0m     \u001b[0moptimizer\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mstep\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m~\\miniconda3\\lib\\site-packages\\torch\\_tensor.py\u001b[0m in \u001b[0;36mbackward\u001b[1;34m(self, gradient, retain_graph, create_graph, inputs)\u001b[0m\n\u001b[0;32m    253\u001b[0m                 \u001b[0mcreate_graph\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mcreate_graph\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    254\u001b[0m                 inputs=inputs)\n\u001b[1;32m--> 255\u001b[1;33m         \u001b[0mtorch\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mautograd\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mbackward\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mgradient\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mretain_graph\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mcreate_graph\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0minputs\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0minputs\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    256\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    257\u001b[0m     \u001b[1;32mdef\u001b[0m \u001b[0mregister_hook\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mhook\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m~\\miniconda3\\lib\\site-packages\\torch\\autograd\\__init__.py\u001b[0m in \u001b[0;36mbackward\u001b[1;34m(tensors, grad_tensors, retain_graph, create_graph, grad_variables, inputs)\u001b[0m\n\u001b[0;32m    145\u001b[0m         \u001b[0mretain_graph\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mcreate_graph\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    146\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 147\u001b[1;33m     Variable._execution_engine.run_backward(\n\u001b[0m\u001b[0;32m    148\u001b[0m         \u001b[0mtensors\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mgrad_tensors_\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mretain_graph\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mcreate_graph\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0minputs\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    149\u001b[0m         allow_unreachable=True, accumulate_grad=True)  # allow_unreachable flag\n",
      "\u001b[1;31mRuntimeError\u001b[0m: CUDA out of memory. Tried to allocate 1.47 GiB (GPU 0; 8.00 GiB total capacity; 5.26 GiB already allocated; 468.12 MiB free; 5.68 GiB reserved in total by PyTorch)"
     ]
    }
   ],
   "source": [
    "blstmModel = BLSTM()\n",
    "torch.cuda.empty_cache()\n",
    "\n",
    "def initWeight(m):\n",
    "    if isinstance(m, torch.nn.Linear) or isinstance(m, torch.nn.LSTM):\n",
    "        torch.nn.init.uniform_(m.weight)\n",
    "        torch.nn.init.ones_(m.bias)\n",
    "        \n",
    "initWeight(blstmModel)\n",
    "\n",
    "optimizer = torch.optim.Adam(blstmModel.parameters(), lr=0.01)\n",
    "lossFn = torch.nn.CrossEntropyLoss()\n",
    "\n",
    "#Needed to use only a subset of the training data due to limited GPU memory availability\n",
    "numItems = int(trainingVectors.size(0)/12)\n",
    "for i in range(500):\n",
    "    optimizer.zero_grad()\n",
    "    #tempOut=torch.as_tensor([item[-1] for item in blstmModel.forward(trainingVectors[:numItems]).tolist()]).cuda()\n",
    "    tempOut = blstmModel.forward(trainingVectors[:numItems])\n",
    "    out = torch.flatten(tempOut, start_dim=0, end_dim=1)\n",
    "    #print(tempOut)\n",
    "    #print(tempOut.shape)\n",
    "#     print(out.shape)\n",
    "#     print(trainingLabels[:numItems].shape)\n",
    "    loss = lossFn(out, trainingLabels[:numItems*952].long())\n",
    "    if i%10 == 0:\n",
    "        print(loss) \n",
    "    loss.backward()\n",
    "    optimizer.step()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "134f5d92-5912-416f-9eaa-4d7d51b40814",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.9.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}