preprocessing.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "from matplotlib import pyplot as plt\n",
    "import seaborn as sns\n",
    "import os\n",
    "from pathlib import Path\n",
    "from sentence_transformers import SentenceTransformer, util\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "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>discourse_text</th>\n",
       "      <th>discourse_type</th>\n",
       "      <th>topics</th>\n",
       "      <th>predictions</th>\n",
       "      <th>context_predictions</th>\n",
       "      <th>augmented_predictions</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>24724</th>\n",
       "      <td>I'm with we should just stay voting for presid...</td>\n",
       "      <td>Position</td>\n",
       "      <td>0</td>\n",
       "      <td>We should keep the Electoral College for the v...</td>\n",
       "      <td>I have a couple reasons for saying that. First...</td>\n",
       "      <td>We should keep the Electoral College for the v...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27275</th>\n",
       "      <td>So having a second opinion is good to have.</td>\n",
       "      <td>Position</td>\n",
       "      <td>7</td>\n",
       "      <td>going to a second opinion is beneficial toward...</td>\n",
       "      <td>going to a second opinion is beneficial toward...</td>\n",
       "      <td>going to a second opinion is beneficial toward...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>101</th>\n",
       "      <td>I disagree with the principal that every stude...</td>\n",
       "      <td>Claim</td>\n",
       "      <td>11</td>\n",
       "      <td>If a principal made every student engage in an...</td>\n",
       "      <td>I disagree that all students must participate ...</td>\n",
       "      <td>I disagree that all students must participate ...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                          discourse_text discourse_type  \\\n",
       "24724  I'm with we should just stay voting for presid...       Position   \n",
       "27275       So having a second opinion is good to have.        Position   \n",
       "101    I disagree with the principal that every stude...          Claim   \n",
       "\n",
       "       topics                                        predictions  \\\n",
       "24724       0  We should keep the Electoral College for the v...   \n",
       "27275       7  going to a second opinion is beneficial toward...   \n",
       "101        11  If a principal made every student engage in an...   \n",
       "\n",
       "                                     context_predictions  \\\n",
       "24724  I have a couple reasons for saying that. First...   \n",
       "27275  going to a second opinion is beneficial toward...   \n",
       "101    I disagree that all students must participate ...   \n",
       "\n",
       "                                   augmented_predictions  \n",
       "24724  We should keep the Electoral College for the v...  \n",
       "27275  going to a second opinion is beneficial toward...  \n",
       "101    I disagree that all students must participate ...  "
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import pandas as pd\n",
    "df = pd.read_csv('data/effective/augmented_predictions_all.csv')\n",
    "df = df[['discourse_text', 'discourse_type', 'topics', 'predictions', 'context_predictions', 'augmented_predictions']]\n",
    "df.sample(3)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "type= {0 : \"electoral_vote_college_president_votes\", \n",
    "       1 : \"emotions_technology_computer_facial_happy\", \n",
    "       2 : \"venus_planet_earth_author_surface\", \n",
    "       3 : \"face_mars_aliens_landform_natural\", \n",
    "       4 : \"online_students_classes_home_school\", \n",
    "       5 : \"community_service_help_kids_think\", \n",
    "       6 : \"summer_project_projects_students_teachers\", \n",
    "       7 : \"advice_multiple_opinions_people_ask\", \n",
    "       8 : \"cars_driverless_car_driver_driving\", \n",
    "       9 : \"cars_car_usage_smog_paris\", \n",
    "       10 : \"phones_cell_phone_school_policy\", \n",
    "       11 : \"extracurricular_activity_activities_students_school\",\n",
    "       12 : \"luke_seagoing_animals_program_cowboys\", \n",
    "       13 : \"average_sports_play_grade_grades\", \n",
    "       14 : \"driving_phone_texting_cell_phones\"}\n",
    "\n",
    "df['topics'] = df['topics'].map(lambda x: type[x])\n",
    "df['discourse_type'] = df['discourse_type'].str.lower()\n",
    "df['len'] = df['discourse_text'].map(lambda x: len(x))\n",
    "df = df[df['len'] < 1024]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "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>discourse_text</th>\n",
       "      <th>discourse_type</th>\n",
       "      <th>topics</th>\n",
       "      <th>predictions</th>\n",
       "      <th>context_predictions</th>\n",
       "      <th>augmented_predictions</th>\n",
       "      <th>len</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>81</th>\n",
       "      <td>The study the ability of humans to read subatl...</td>\n",
       "      <td>evidence</td>\n",
       "      <td>emotions_technology_computer_facial_happy</td>\n",
       "      <td>Every living thing on planet earth has emotion...</td>\n",
       "      <td>Cars in most countrioes are the leading pollut...</td>\n",
       "      <td>In today's world, students are filled with man...</td>\n",
       "      <td>1449</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>86</th>\n",
       "      <td>The Mona lisa demonstration really intended to...</td>\n",
       "      <td>evidence</td>\n",
       "      <td>emotions_technology_computer_facial_happy</td>\n",
       "      <td>According to the article \"Making Mona Lisa Smi...</td>\n",
       "      <td>With this in our cumputers it can read the emo...</td>\n",
       "      <td>According to the article \"Making Mona Lisa Smi...</td>\n",
       "      <td>1165</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>174</th>\n",
       "      <td>Attention !!! to all the residents of this com...</td>\n",
       "      <td>evidence</td>\n",
       "      <td>cars_car_usage_smog_paris</td>\n",
       "      <td>Cars have always been apart of American cultur...</td>\n",
       "      <td>Cars in most countrioes are the leading pollut...</td>\n",
       "      <td>There's some of us who suffer from stress and ...</td>\n",
       "      <td>1974</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>288</th>\n",
       "      <td>Luke joined the program. He decided to because...</td>\n",
       "      <td>evidence</td>\n",
       "      <td>luke_seagoing_animals_program_cowboys</td>\n",
       "      <td>Hi, I'm Luke Bomberger. I was a Seagoing Cowbo...</td>\n",
       "      <td>Cars in most countrioes are the leading pollut...</td>\n",
       "      <td>The Seagoing Cowboys program is a fun experien...</td>\n",
       "      <td>1039</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>291</th>\n",
       "      <td>Your frontalis pars lateralis muscle above you...</td>\n",
       "      <td>evidence</td>\n",
       "      <td>emotions_technology_computer_facial_happy</td>\n",
       "      <td>like its says in paragraph 6 \"Imagine a comput...</td>\n",
       "      <td>While there are many positives to this technol...</td>\n",
       "      <td>The Facial Action Coding System enables comput...</td>\n",
       "      <td>1109</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27101</th>\n",
       "      <td>In conclusion, considering the idea for studen...</td>\n",
       "      <td>concluding statement</td>\n",
       "      <td>online_students_classes_home_school</td>\n",
       "      <td>Taking classes online and learning from home b...</td>\n",
       "      <td>but the parent may have to go to work or run e...</td>\n",
       "      <td>Online or video conferencing is seen by school...</td>\n",
       "      <td>1125</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27107</th>\n",
       "      <td>Different people have different answers depend...</td>\n",
       "      <td>lead</td>\n",
       "      <td>advice_multiple_opinions_people_ask</td>\n",
       "      <td>As science is an ever-budding field by it's ow...</td>\n",
       "      <td>you see secretly i was showing examples of peo...</td>\n",
       "      <td>As in the physical world, inertia carries idea...</td>\n",
       "      <td>1075</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27110</th>\n",
       "      <td>You should always remember that perpendiculari...</td>\n",
       "      <td>evidence</td>\n",
       "      <td>advice_multiple_opinions_people_ask</td>\n",
       "      <td>it is on the brink of catastrophe.</td>\n",
       "      <td>you see secretly i was showing examples of peo...</td>\n",
       "      <td>Assuming that the exploration of Venus does ev...</td>\n",
       "      <td>1609</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27114</th>\n",
       "      <td>Always remember that the perpendicularity of t...</td>\n",
       "      <td>concluding statement</td>\n",
       "      <td>advice_multiple_opinions_people_ask</td>\n",
       "      <td>it is on the brink of catastrophe.</td>\n",
       "      <td>you see secretly i was showing examples of peo...</td>\n",
       "      <td>I'm a scientist at NASA and the Face, a lot of...</td>\n",
       "      <td>1142</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27333</th>\n",
       "      <td>That is so important for me and my parents and...</td>\n",
       "      <td>evidence</td>\n",
       "      <td>online_students_classes_home_school</td>\n",
       "      <td>School can help students achieve a greter goal...</td>\n",
       "      <td>Most students dislike the idea of waking up in...</td>\n",
       "      <td>Its important for you to attend class in a sch...</td>\n",
       "      <td>1191</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>315 rows × 7 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                          discourse_text  \\\n",
       "81     The study the ability of humans to read subatl...   \n",
       "86     The Mona lisa demonstration really intended to...   \n",
       "174    Attention !!! to all the residents of this com...   \n",
       "288    Luke joined the program. He decided to because...   \n",
       "291    Your frontalis pars lateralis muscle above you...   \n",
       "...                                                  ...   \n",
       "27101  In conclusion, considering the idea for studen...   \n",
       "27107  Different people have different answers depend...   \n",
       "27110  You should always remember that perpendiculari...   \n",
       "27114  Always remember that the perpendicularity of t...   \n",
       "27333  That is so important for me and my parents and...   \n",
       "\n",
       "             discourse_type                                     topics  \\\n",
       "81                 evidence  emotions_technology_computer_facial_happy   \n",
       "86                 evidence  emotions_technology_computer_facial_happy   \n",
       "174                evidence                  cars_car_usage_smog_paris   \n",
       "288                evidence      luke_seagoing_animals_program_cowboys   \n",
       "291                evidence  emotions_technology_computer_facial_happy   \n",
       "...                     ...                                        ...   \n",
       "27101  concluding statement        online_students_classes_home_school   \n",
       "27107                  lead        advice_multiple_opinions_people_ask   \n",
       "27110              evidence        advice_multiple_opinions_people_ask   \n",
       "27114  concluding statement        advice_multiple_opinions_people_ask   \n",
       "27333              evidence        online_students_classes_home_school   \n",
       "\n",
       "                                             predictions  \\\n",
       "81     Every living thing on planet earth has emotion...   \n",
       "86     According to the article \"Making Mona Lisa Smi...   \n",
       "174    Cars have always been apart of American cultur...   \n",
       "288    Hi, I'm Luke Bomberger. I was a Seagoing Cowbo...   \n",
       "291    like its says in paragraph 6 \"Imagine a comput...   \n",
       "...                                                  ...   \n",
       "27101  Taking classes online and learning from home b...   \n",
       "27107  As science is an ever-budding field by it's ow...   \n",
       "27110                it is on the brink of catastrophe.    \n",
       "27114                it is on the brink of catastrophe.    \n",
       "27333  School can help students achieve a greter goal...   \n",
       "\n",
       "                                     context_predictions  \\\n",
       "81     Cars in most countrioes are the leading pollut...   \n",
       "86     With this in our cumputers it can read the emo...   \n",
       "174    Cars in most countrioes are the leading pollut...   \n",
       "288    Cars in most countrioes are the leading pollut...   \n",
       "291    While there are many positives to this technol...   \n",
       "...                                                  ...   \n",
       "27101  but the parent may have to go to work or run e...   \n",
       "27107  you see secretly i was showing examples of peo...   \n",
       "27110  you see secretly i was showing examples of peo...   \n",
       "27114  you see secretly i was showing examples of peo...   \n",
       "27333  Most students dislike the idea of waking up in...   \n",
       "\n",
       "                                   augmented_predictions   len  \n",
       "81     In today's world, students are filled with man...  1449  \n",
       "86     According to the article \"Making Mona Lisa Smi...  1165  \n",
       "174    There's some of us who suffer from stress and ...  1974  \n",
       "288    The Seagoing Cowboys program is a fun experien...  1039  \n",
       "291    The Facial Action Coding System enables comput...  1109  \n",
       "...                                                  ...   ...  \n",
       "27101  Online or video conferencing is seen by school...  1125  \n",
       "27107  As in the physical world, inertia carries idea...  1075  \n",
       "27110  Assuming that the exploration of Venus does ev...  1609  \n",
       "27114  I'm a scientist at NASA and the Face, a lot of...  1142  \n",
       "27333  Its important for you to attend class in a sch...  1191  \n",
       "\n",
       "[315 rows x 7 columns]"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test = df.copy()\n",
    "test['len'] = test['discourse_text'].map(lambda x: len(x))\n",
    "test[test['len'] < 1024]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "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>discourse_text</th>\n",
       "      <th>discourse_type</th>\n",
       "      <th>topics</th>\n",
       "      <th>predictions</th>\n",
       "      <th>context_predictions</th>\n",
       "      <th>augmented_predictions</th>\n",
       "      <th>len</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Hi, i'm Isaac, i'm going to be writing about h...</td>\n",
       "      <td>lead</td>\n",
       "      <td>face_mars_aliens_landform_natural</td>\n",
       "      <td>The curiousity about Mars has been going on fo...</td>\n",
       "      <td>In 1976, Viking 1 took photos of Mars and foun...</td>\n",
       "      <td>The curiousity about Mars has been going on fo...</td>\n",
       "      <td>317</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>On my perspective, I think that the face is a ...</td>\n",
       "      <td>position</td>\n",
       "      <td>face_mars_aliens_landform_natural</td>\n",
       "      <td>However NASA has provided evidence that proves...</td>\n",
       "      <td>When Mars Orbiter Camera team snap pictures it...</td>\n",
       "      <td>However NASA has provided evidence that proves...</td>\n",
       "      <td>210</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>I think that the face is a natural landform be...</td>\n",
       "      <td>claim</td>\n",
       "      <td>face_mars_aliens_landform_natural</td>\n",
       "      <td>However NASA has provided evidence that proves...</td>\n",
       "      <td>However NASA has provided evidence that proves...</td>\n",
       "      <td>However NASA has provided evidence that proves...</td>\n",
       "      <td>105</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>If life was on Mars, we would know by now. The...</td>\n",
       "      <td>evidence</td>\n",
       "      <td>face_mars_aliens_landform_natural</td>\n",
       "      <td>In the end, the landform was found to be a Mar...</td>\n",
       "      <td>In 1976, Viking 1 took photos of Mars and foun...</td>\n",
       "      <td>In the end, the landform was found to be a Mar...</td>\n",
       "      <td>362</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>People thought that the face was formed by ali...</td>\n",
       "      <td>counterclaim</td>\n",
       "      <td>face_mars_aliens_landform_natural</td>\n",
       "      <td>Contrary to popular belief, there is nothing u...</td>\n",
       "      <td>In 1976, Viking 1 took photos of Mars and foun...</td>\n",
       "      <td>Contrary to popular belief, there is nothing u...</td>\n",
       "      <td>101</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27434</th>\n",
       "      <td>For many people they don't like only asking on...</td>\n",
       "      <td>claim</td>\n",
       "      <td>advice_multiple_opinions_people_ask</td>\n",
       "      <td>Because one person might have better advice th...</td>\n",
       "      <td>Advice is an opinion, opinions have bias in th...</td>\n",
       "      <td>Advice is an opinion, opinions have bias in th...</td>\n",
       "      <td>126</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27435</th>\n",
       "      <td>also people have different views and opinions ...</td>\n",
       "      <td>claim</td>\n",
       "      <td>advice_multiple_opinions_people_ask</td>\n",
       "      <td>people have different views.</td>\n",
       "      <td>Advice is an opinion, opinions have bias in th...</td>\n",
       "      <td>Everyone thinks differently and has their own ...</td>\n",
       "      <td>57</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27436</th>\n",
       "      <td>Advice is something that can impact a persons ...</td>\n",
       "      <td>position</td>\n",
       "      <td>advice_multiple_opinions_people_ask</td>\n",
       "      <td>Good advice is something that will help you fi...</td>\n",
       "      <td>Advice is an opinion, opinions have bias in th...</td>\n",
       "      <td>Advice is an opinion, opinions have bias in th...</td>\n",
       "      <td>121</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27437</th>\n",
       "      <td>someone can use everything that many people sa...</td>\n",
       "      <td>evidence</td>\n",
       "      <td>advice_multiple_opinions_people_ask</td>\n",
       "      <td>The method of asking many people for their adv...</td>\n",
       "      <td>Now this reason would probably apply to someon...</td>\n",
       "      <td>Advice is an opinion, opinions have bias in th...</td>\n",
       "      <td>439</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27438</th>\n",
       "      <td>In conclusion asking for an opinion can be ben...</td>\n",
       "      <td>concluding statement</td>\n",
       "      <td>advice_multiple_opinions_people_ask</td>\n",
       "      <td>When you ask for someones opinion you should l...</td>\n",
       "      <td>Now this reason would probably apply to someon...</td>\n",
       "      <td>seeking advice is strongly advised. Asking for...</td>\n",
       "      <td>67</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>27123 rows × 7 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                          discourse_text  \\\n",
       "0      Hi, i'm Isaac, i'm going to be writing about h...   \n",
       "1      On my perspective, I think that the face is a ...   \n",
       "2      I think that the face is a natural landform be...   \n",
       "3      If life was on Mars, we would know by now. The...   \n",
       "4      People thought that the face was formed by ali...   \n",
       "...                                                  ...   \n",
       "27434  For many people they don't like only asking on...   \n",
       "27435  also people have different views and opinions ...   \n",
       "27436  Advice is something that can impact a persons ...   \n",
       "27437  someone can use everything that many people sa...   \n",
       "27438  In conclusion asking for an opinion can be ben...   \n",
       "\n",
       "             discourse_type                               topics  \\\n",
       "0                      lead    face_mars_aliens_landform_natural   \n",
       "1                  position    face_mars_aliens_landform_natural   \n",
       "2                     claim    face_mars_aliens_landform_natural   \n",
       "3                  evidence    face_mars_aliens_landform_natural   \n",
       "4              counterclaim    face_mars_aliens_landform_natural   \n",
       "...                     ...                                  ...   \n",
       "27434                 claim  advice_multiple_opinions_people_ask   \n",
       "27435                 claim  advice_multiple_opinions_people_ask   \n",
       "27436              position  advice_multiple_opinions_people_ask   \n",
       "27437              evidence  advice_multiple_opinions_people_ask   \n",
       "27438  concluding statement  advice_multiple_opinions_people_ask   \n",
       "\n",
       "                                             predictions  \\\n",
       "0      The curiousity about Mars has been going on fo...   \n",
       "1      However NASA has provided evidence that proves...   \n",
       "2      However NASA has provided evidence that proves...   \n",
       "3      In the end, the landform was found to be a Mar...   \n",
       "4      Contrary to popular belief, there is nothing u...   \n",
       "...                                                  ...   \n",
       "27434  Because one person might have better advice th...   \n",
       "27435                      people have different views.    \n",
       "27436  Good advice is something that will help you fi...   \n",
       "27437  The method of asking many people for their adv...   \n",
       "27438  When you ask for someones opinion you should l...   \n",
       "\n",
       "                                     context_predictions  \\\n",
       "0      In 1976, Viking 1 took photos of Mars and foun...   \n",
       "1      When Mars Orbiter Camera team snap pictures it...   \n",
       "2      However NASA has provided evidence that proves...   \n",
       "3      In 1976, Viking 1 took photos of Mars and foun...   \n",
       "4      In 1976, Viking 1 took photos of Mars and foun...   \n",
       "...                                                  ...   \n",
       "27434  Advice is an opinion, opinions have bias in th...   \n",
       "27435  Advice is an opinion, opinions have bias in th...   \n",
       "27436  Advice is an opinion, opinions have bias in th...   \n",
       "27437  Now this reason would probably apply to someon...   \n",
       "27438  Now this reason would probably apply to someon...   \n",
       "\n",
       "                                   augmented_predictions  len  \n",
       "0      The curiousity about Mars has been going on fo...  317  \n",
       "1      However NASA has provided evidence that proves...  210  \n",
       "2      However NASA has provided evidence that proves...  105  \n",
       "3      In the end, the landform was found to be a Mar...  362  \n",
       "4      Contrary to popular belief, there is nothing u...  101  \n",
       "...                                                  ...  ...  \n",
       "27434  Advice is an opinion, opinions have bias in th...  126  \n",
       "27435  Everyone thinks differently and has their own ...   57  \n",
       "27436  Advice is an opinion, opinions have bias in th...  121  \n",
       "27437  Advice is an opinion, opinions have bias in th...  439  \n",
       "27438  seeking advice is strongly advised. Asking for...   67  \n",
       "\n",
       "[27123 rows x 7 columns]"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Axes: xlabel='len', ylabel='Density'>"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.kdeplot(test['len'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#sum(df['augmented_predictions'].str.contains('\\u00a0'))\n",
    "#sum(df['augmented_predictions'].str.contains('\\xa0'))\n",
    "sum(df['augmented_predictions'].str.contains('\\n'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "df.to_csv('data/effective/augmented_predictions_all.csv')"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "argue-better",
   "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": "3.8.16"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}