LAB 1 - Data Preparation.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "adbf511d-14d2-4b64-893c-de6234aff00e",
   "metadata": {},
   "source": [
    "<img src=\"static/event-image-small.jpg\" alt=\"drawing\" width=\"800\"/> \n",
    "\n",
    "# **LAB 1: DATA PREPARATION**\n",
    "\n",
    "In this notebook, we will explore how do prepare data for downstream NLP tasks.  \n",
    "\n",
    "A non-exhaustive list tasks related to NLP and Generative AI that can benefit from this data\n",
    "preparation is:\n",
    "\n",
    "- Question & Answer\n",
    "- Text Summarization\n",
    "- Instruct Tuning\n",
    "- Human-Bot Conversations\n",
    "- Continued PreTraining\n",
    "\n",
    "\n",
    "For each of these tasks, there are subtle differences in how one might want to prepare the data.\n",
    "Some of the key steps one might use here for data prep are also applicable to \n",
    "applying guardrails to LLMs (e.g., `Profanity Check` and `Toxicity Detection`)\n",
    "\n",
    "More on guardrails in Lab #5, but for now let's import our required python \n",
    "libraries and get started!\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "31a97725-7a33-4caa-8fd3-5f55de7503e7",
   "metadata": {},
   "source": [
    "# Import Required Dependencies "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "754aaf1a-6ac3-4ec9-af02-7975a93ce85d",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "import warnings\n",
    "warnings.filterwarnings('ignore')\n",
    "\n",
    "# set flag for training environment\n",
    "TRAINING = True\n",
    "\n",
    "import torch\n",
    "from transformers import GPT2Tokenizer, GPT2LMHeadModel\n",
    "from torch.utils.data import Dataset, DataLoader\n",
    "import pandas as pd \n",
    "import numpy as np\n",
    "\n",
    "import matplotlib"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "afffc81d",
   "metadata": {},
   "source": [
    "# Dataset\n",
    "\n",
    "For most of this training we will focus on a public dataset that focuses on text content\n",
    "and metadata for LinkedIn posts during 2021. These posts were collected from the\n",
    "internet and correspond to a wide array of identified \"influencers\".\n",
    "\n",
    "The data set can be found at: https://www.kaggle.com/datasets/shreyasajal/linkedin-influencers-data\n",
    "\n",
    "Noticeably, this dataset has seen a spike of new downloads in the last few weeks, \n",
    "perhaps related to an increased awareness of possible llm applications\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4cdb0c26",
   "metadata": {},
   "source": [
    "\n",
    "<img src=\"static/data-downloads.png\" alt=\"drawing\" width=\"800\"/> "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "20b331d7",
   "metadata": {},
   "source": [
    "# Import Dataset\n",
    "\n",
    "For convenience, H2O.ai has downloaded this dataset and placed it on a public\n",
    "bucket on AWS S3. Let's read it in and take a look!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "a5edece6",
   "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>name</th>\n",
       "      <th>headline</th>\n",
       "      <th>location</th>\n",
       "      <th>followers</th>\n",
       "      <th>connections</th>\n",
       "      <th>about</th>\n",
       "      <th>time_spent</th>\n",
       "      <th>content</th>\n",
       "      <th>content_links</th>\n",
       "      <th>media_type</th>\n",
       "      <th>media_url</th>\n",
       "      <th>num_hashtags</th>\n",
       "      <th>hashtag_followers</th>\n",
       "      <th>hashtags</th>\n",
       "      <th>reactions</th>\n",
       "      <th>comments</th>\n",
       "      <th>views</th>\n",
       "      <th>votes</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>34007</th>\n",
       "      <td>Simon Sinek</td>\n",
       "      <td>Optimist and Author at Simon Sinek Inc.</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4206024.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Simon Sinek is an unshakable optimist who beli...</td>\n",
       "      <td>4 years ago</td>\n",
       "      <td>Igniter of the Year 2016. Well I know that I'm...</td>\n",
       "      <td>[]</td>\n",
       "      <td>image</td>\n",
       "      <td>['https://media-exp1.licdn.com/dms/image/C5622...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>[]</td>\n",
       "      <td>4005</td>\n",
       "      <td>93</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>34008</th>\n",
       "      <td>Simon Sinek</td>\n",
       "      <td>Optimist and Author at Simon Sinek Inc.</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4206024.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Simon Sinek is an unshakable optimist who beli...</td>\n",
       "      <td>4 years ago</td>\n",
       "      <td>Executives who prioritize the shareholder are ...</td>\n",
       "      <td>[]</td>\n",
       "      <td>video</td>\n",
       "      <td>[]</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>[]</td>\n",
       "      <td>1698</td>\n",
       "      <td>74</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>34009</th>\n",
       "      <td>Simon Sinek</td>\n",
       "      <td>Optimist and Author at Simon Sinek Inc.</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4206024.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Simon Sinek is an unshakable optimist who beli...</td>\n",
       "      <td>4 years ago</td>\n",
       "      <td>Like many, I too have been reflecting as we ne...</td>\n",
       "      <td>[]</td>\n",
       "      <td>video</td>\n",
       "      <td>[]</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>[]</td>\n",
       "      <td>661</td>\n",
       "      <td>59</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>34010</th>\n",
       "      <td>Simon Sinek</td>\n",
       "      <td>Optimist and Author at Simon Sinek Inc.</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4206024.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Simon Sinek is an unshakable optimist who beli...</td>\n",
       "      <td>4 years ago</td>\n",
       "      <td>If you say \"customer first\" that means your em...</td>\n",
       "      <td>[]</td>\n",
       "      <td>video</td>\n",
       "      <td>[]</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>[]</td>\n",
       "      <td>766</td>\n",
       "      <td>35</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>34011</th>\n",
       "      <td>Simon Sinek</td>\n",
       "      <td>Optimist and Author at Simon Sinek Inc.</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4206024.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Simon Sinek is an unshakable optimist who beli...</td>\n",
       "      <td>4 years ago</td>\n",
       "      <td>The small work hard to serve themselves in a b...</td>\n",
       "      <td>[]</td>\n",
       "      <td>NaN</td>\n",
       "      <td>[]</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>[]</td>\n",
       "      <td>789</td>\n",
       "      <td>23</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              name                                 headline location  \\\n",
       "34007  Simon Sinek  Optimist and Author at Simon Sinek Inc.      NaN   \n",
       "34008  Simon Sinek  Optimist and Author at Simon Sinek Inc.      NaN   \n",
       "34009  Simon Sinek  Optimist and Author at Simon Sinek Inc.      NaN   \n",
       "34010  Simon Sinek  Optimist and Author at Simon Sinek Inc.      NaN   \n",
       "34011  Simon Sinek  Optimist and Author at Simon Sinek Inc.      NaN   \n",
       "\n",
       "       followers connections  \\\n",
       "34007  4206024.0         NaN   \n",
       "34008  4206024.0         NaN   \n",
       "34009  4206024.0         NaN   \n",
       "34010  4206024.0         NaN   \n",
       "34011  4206024.0         NaN   \n",
       "\n",
       "                                                   about   time_spent  \\\n",
       "34007  Simon Sinek is an unshakable optimist who beli...  4 years ago   \n",
       "34008  Simon Sinek is an unshakable optimist who beli...  4 years ago   \n",
       "34009  Simon Sinek is an unshakable optimist who beli...  4 years ago   \n",
       "34010  Simon Sinek is an unshakable optimist who beli...  4 years ago   \n",
       "34011  Simon Sinek is an unshakable optimist who beli...  4 years ago   \n",
       "\n",
       "                                                 content content_links  \\\n",
       "34007  Igniter of the Year 2016. Well I know that I'm...            []   \n",
       "34008  Executives who prioritize the shareholder are ...            []   \n",
       "34009  Like many, I too have been reflecting as we ne...            []   \n",
       "34010  If you say \"customer first\" that means your em...            []   \n",
       "34011  The small work hard to serve themselves in a b...            []   \n",
       "\n",
       "      media_type                                          media_url  \\\n",
       "34007      image  ['https://media-exp1.licdn.com/dms/image/C5622...   \n",
       "34008      video                                                 []   \n",
       "34009      video                                                 []   \n",
       "34010      video                                                 []   \n",
       "34011        NaN                                                 []   \n",
       "\n",
       "       num_hashtags  hashtag_followers hashtags  reactions  comments  views  \\\n",
       "34007             0                  0       []       4005        93    NaN   \n",
       "34008             0                  0       []       1698        74    NaN   \n",
       "34009             0                  0       []        661        59    NaN   \n",
       "34010             0                  0       []        766        35    NaN   \n",
       "34011             0                  0       []        789        23    NaN   \n",
       "\n",
       "      votes  \n",
       "34007   NaN  \n",
       "34008   NaN  \n",
       "34009   NaN  \n",
       "34010   NaN  \n",
       "34011   NaN  "
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Read in from public S3 Location\n",
    "df = pd.read_csv(\"s3a://h2o-public-test-data/generative-ai/influencers_data.csv\")\n",
    "df.tail()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "4bee2b35",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 34012 entries, 0 to 34011\n",
      "Data columns (total 18 columns):\n",
      " #   Column             Non-Null Count  Dtype  \n",
      "---  ------             --------------  -----  \n",
      " 0   name               34012 non-null  object \n",
      " 1   headline           34012 non-null  object \n",
      " 2   location           31740 non-null  object \n",
      " 3   followers          33970 non-null  float64\n",
      " 4   connections        25713 non-null  object \n",
      " 5   about              34012 non-null  object \n",
      " 6   time_spent         34011 non-null  object \n",
      " 7   content            31996 non-null  object \n",
      " 8   content_links      34012 non-null  object \n",
      " 9   media_type         26779 non-null  object \n",
      " 10  media_url          34012 non-null  object \n",
      " 11  num_hashtags       34012 non-null  int64  \n",
      " 12  hashtag_followers  34012 non-null  int64  \n",
      " 13  hashtags           34012 non-null  object \n",
      " 14  reactions          34012 non-null  int64  \n",
      " 15  comments           34012 non-null  int64  \n",
      " 16  views              0 non-null      float64\n",
      " 17  votes              86 non-null     object \n",
      "dtypes: float64(2), int64(4), object(12)\n",
      "memory usage: 4.7+ MB\n"
     ]
    }
   ],
   "source": [
    "df.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "61711ff2",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(34012, 18)"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b058c6c9",
   "metadata": {},
   "source": [
    "There are approximately ~70 LinkedIn influencers in this dataset. Let's take a look at\n",
    "how many posts we have data on for each influencer."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "2f9d46d0",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Axes: ylabel='name'>"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x1200 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# count how many records we have for influencers\n",
    "df.name.value_counts().sort_values().plot(kind='barh', figsize=(8,12))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "06f06ead",
   "metadata": {},
   "source": [
    "# Create Small Sample DataFrame"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "3705cf8f",
   "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>name</th>\n",
       "      <th>headline</th>\n",
       "      <th>content</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>19353</th>\n",
       "      <td>Rachel Jones</td>\n",
       "      <td>Communications Consultant at Emerson Collective</td>\n",
       "      <td>Professor  Mmaki Jantjies (PhD) , here's a bri...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28604</th>\n",
       "      <td>Shelly Palmer</td>\n",
       "      <td>CEO at The Palmer Group</td>\n",
       "      <td>Under the measure, workers must be designated ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11363</th>\n",
       "      <td>Lynne Everatt</td>\n",
       "      <td>Author of The 5-Minute Recharge and Acts of Fr...</td>\n",
       "      <td>Today we’re better off in a material sense, bu...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30744</th>\n",
       "      <td>Andrew Yang</td>\n",
       "      <td>Candidate for NYC Mayor in 2021 (D) at Yangfor...</td>\n",
       "      <td>It is impossible to overstate the positive imp...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11276</th>\n",
       "      <td>Lynne Everatt</td>\n",
       "      <td>Author of The 5-Minute Recharge and Acts of Fr...</td>\n",
       "      <td>Want to give your relationship a booster shot?...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                name                                           headline  \\\n",
       "19353   Rachel Jones    Communications Consultant at Emerson Collective   \n",
       "28604  Shelly Palmer                            CEO at The Palmer Group   \n",
       "11363  Lynne Everatt  Author of The 5-Minute Recharge and Acts of Fr...   \n",
       "30744    Andrew Yang  Candidate for NYC Mayor in 2021 (D) at Yangfor...   \n",
       "11276  Lynne Everatt  Author of The 5-Minute Recharge and Acts of Fr...   \n",
       "\n",
       "                                                 content  \n",
       "19353  Professor  Mmaki Jantjies (PhD) , here's a bri...  \n",
       "28604  Under the measure, workers must be designated ...  \n",
       "11363  Today we’re better off in a material sense, bu...  \n",
       "30744  It is impossible to overstate the positive imp...  \n",
       "11276  Want to give your relationship a booster shot?...  "
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sample_df = df.sample(100)[['name', 'headline', 'content']]\n",
    "\n",
    "# drop records with missing content\n",
    "sample_df = sample_df.dropna()\n",
    "\n",
    "sample_df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ebc27478",
   "metadata": {},
   "source": [
    "# Custom Functions\n",
    "\n",
    "As anyone who has done any sort of data engineering, or even just basic data prep\n",
    "before modeling knows, setting up custom functions that are modular and repeatable\n",
    "is not only valuable it's best practice\n",
    "\n",
    "Here we introduce an example template custom function that can be used for our \n",
    "purposes as well as later on when we introduce `H2O LLM DataStudio`\n",
    "\n",
    "\n",
    "```\n",
    "def custom_function(df, text_columns):\n",
    "    \n",
    "    # For text cleaning based functions use: \n",
    "    for column in text_columns: \n",
    "        df[column] = # add change code here \n",
    "    \n",
    "    #For an example of how to filter the df see number_proportion_filter.py or numeric_filer.py \n",
    "    \n",
    "    return df\n",
    "```\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "127f2125",
   "metadata": {},
   "source": [
    "# 🙈 Profanity Check\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "f31facd7",
   "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>name</th>\n",
       "      <th>headline</th>\n",
       "      <th>content</th>\n",
       "      <th>profanity</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Rachel Jones</td>\n",
       "      <td>Communications Consultant at Emerson Collective</td>\n",
       "      <td>Professor  Mmaki Jantjies (PhD) , here's a bri...</td>\n",
       "      <td>0.014745</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Shelly Palmer</td>\n",
       "      <td>CEO at The Palmer Group</td>\n",
       "      <td>Under the measure, workers must be designated ...</td>\n",
       "      <td>0.027932</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Lynne Everatt</td>\n",
       "      <td>Author of The 5-Minute Recharge and Acts of Fr...</td>\n",
       "      <td>Today we’re better off in a material sense, bu...</td>\n",
       "      <td>0.020130</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Andrew Yang</td>\n",
       "      <td>Candidate for NYC Mayor in 2021 (D) at Yangfor...</td>\n",
       "      <td>It is impossible to overstate the positive imp...</td>\n",
       "      <td>0.012047</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Lynne Everatt</td>\n",
       "      <td>Author of The 5-Minute Recharge and Acts of Fr...</td>\n",
       "      <td>Want to give your relationship a booster shot?...</td>\n",
       "      <td>0.101273</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            name                                           headline  \\\n",
       "0   Rachel Jones    Communications Consultant at Emerson Collective   \n",
       "1  Shelly Palmer                            CEO at The Palmer Group   \n",
       "2  Lynne Everatt  Author of The 5-Minute Recharge and Acts of Fr...   \n",
       "3    Andrew Yang  Candidate for NYC Mayor in 2021 (D) at Yangfor...   \n",
       "4  Lynne Everatt  Author of The 5-Minute Recharge and Acts of Fr...   \n",
       "\n",
       "                                             content  profanity  \n",
       "0  Professor  Mmaki Jantjies (PhD) , here's a bri...   0.014745  \n",
       "1  Under the measure, workers must be designated ...   0.027932  \n",
       "2  Today we’re better off in a material sense, bu...   0.020130  \n",
       "3  It is impossible to overstate the positive imp...   0.012047  \n",
       "4  Want to give your relationship a booster shot?...   0.101273  "
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from profanity_check import predict_prob\n",
    "\n",
    "threshold = 0.9\n",
    "\n",
    "sample_df[\"profanity\"] = predict_prob(sample_df[\"content\"])\n",
    "\n",
    "sample_df = sample_df[sample_df[\"profanity\"] < threshold]\n",
    "sample_df = sample_df.reset_index(drop=True)\n",
    "sample_df.head()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "d5ba6804",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Axes: >"
      ]
     },
     "execution_count": 8,
     "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": [
    "sample_df['profanity'].hist()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4304a58a",
   "metadata": {},
   "source": [
    "# 🧐 Text Quality Check\n",
    "\n",
    "## Flesch Grade reading level\n",
    "\n",
    "Depending on the task at hand, sometimes it's advantages to evaluate text and\n",
    "language based on it's estimated reading level. One popular way to do this \n",
    "is with the Flesch-Kincaid grade level.  This is a grade formula in that a score\n",
    "of 9.3 means that a ninth grader would be able to read the document.\n",
    "\n",
    "The wikipedia article on this technique does a fair job explaining it:\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "63b8e3be",
   "metadata": {},
   "source": [
    "<img src=\"static/flesch-kincade-scale.png\" alt=\"drawing\" width=\"800\"/> "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1029ea35",
   "metadata": {},
   "source": [
    "Using the `textstat` python linvrary, we can approximate the reading level and\n",
    "set our thresholds for what level we find ideal for our downstream NLP task.\n",
    "\n",
    "In this case, we will limit the reading level between `2` and `10`, but many\n",
    "other ranges are certainly possible"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "ed85881a",
   "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>name</th>\n",
       "      <th>headline</th>\n",
       "      <th>content</th>\n",
       "      <th>profanity</th>\n",
       "      <th>flesch_grade</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Lynne Everatt</td>\n",
       "      <td>Author of The 5-Minute Recharge and Acts of Fr...</td>\n",
       "      <td>Today we’re better off in a material sense, bu...</td>\n",
       "      <td>0.020130</td>\n",
       "      <td>7.2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Andrew Yang</td>\n",
       "      <td>Candidate for NYC Mayor in 2021 (D) at Yangfor...</td>\n",
       "      <td>It is impossible to overstate the positive imp...</td>\n",
       "      <td>0.012047</td>\n",
       "      <td>8.2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Lynne Everatt</td>\n",
       "      <td>Author of The 5-Minute Recharge and Acts of Fr...</td>\n",
       "      <td>Want to give your relationship a booster shot?...</td>\n",
       "      <td>0.101273</td>\n",
       "      <td>5.4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>Richard Branson</td>\n",
       "      <td>Founder at Virgin Group</td>\n",
       "      <td>It’s been a pleasure working with Yanik Silver...</td>\n",
       "      <td>0.116892</td>\n",
       "      <td>4.2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>AJ Wilcox</td>\n",
       "      <td>LinkedIn Ads Fanatic and Host of The LinkedIn ...</td>\n",
       "      <td>Got to be a guest on the \"Inbound and Down\" po...</td>\n",
       "      <td>0.048126</td>\n",
       "      <td>8.1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              name                                           headline  \\\n",
       "2    Lynne Everatt  Author of The 5-Minute Recharge and Acts of Fr...   \n",
       "3      Andrew Yang  Candidate for NYC Mayor in 2021 (D) at Yangfor...   \n",
       "4    Lynne Everatt  Author of The 5-Minute Recharge and Acts of Fr...   \n",
       "5  Richard Branson                            Founder at Virgin Group   \n",
       "8        AJ Wilcox  LinkedIn Ads Fanatic and Host of The LinkedIn ...   \n",
       "\n",
       "                                             content  profanity  flesch_grade  \n",
       "2  Today we’re better off in a material sense, bu...   0.020130           7.2  \n",
       "3  It is impossible to overstate the positive imp...   0.012047           8.2  \n",
       "4  Want to give your relationship a booster shot?...   0.101273           5.4  \n",
       "5  It’s been a pleasure working with Yanik Silver...   0.116892           4.2  \n",
       "8  Got to be a guest on the \"Inbound and Down\" po...   0.048126           8.1  "
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import textstat \n",
    "\n",
    "min_grade_level = 2\n",
    "max_grade_level = 10\n",
    "\n",
    "def check_quality(df, text_column):\n",
    "    df[\"flesch_grade\"] = df[text_column].apply(textstat.flesch_kincaid_grade)\n",
    "    return df\n",
    "\n",
    "cleaned = check_quality(sample_df, 'content')\n",
    "\n",
    "# Filter to only those records with content between 2 and 10 grade level\n",
    "cleaned = cleaned[(cleaned[\"flesch_grade\"] >= min_grade_level) & (cleaned[\"flesch_grade\"] <= max_grade_level)]\n",
    "cleaned.head()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "c57c9008",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Axes: >"
      ]
     },
     "execution_count": 10,
     "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": [
    "sample_df['flesch_grade'].hist()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0293bb09",
   "metadata": {},
   "source": [
    "# Run Preparation Pipeline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "5e1288eb",
   "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>name</th>\n",
       "      <th>headline</th>\n",
       "      <th>about</th>\n",
       "      <th>content</th>\n",
       "      <th>reactions</th>\n",
       "      <th>profanity</th>\n",
       "      <th>flesch_grade</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Nicholas Wyman</td>\n",
       "      <td>CEO IWSI Group</td>\n",
       "      <td>Nicholas Wyman for the past 25 years has shone...</td>\n",
       "      <td>National disability advocate  Sara Hart Weir, ...</td>\n",
       "      <td>11</td>\n",
       "      <td>0.018358</td>\n",
       "      <td>8.8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Nicholas Wyman</td>\n",
       "      <td>CEO IWSI Group</td>\n",
       "      <td>Nicholas Wyman for the past 25 years has shone...</td>\n",
       "      <td>Exploring in this months Talent Management &amp; H...</td>\n",
       "      <td>44</td>\n",
       "      <td>0.008754</td>\n",
       "      <td>8.8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Nicholas Wyman</td>\n",
       "      <td>CEO IWSI Group</td>\n",
       "      <td>Nicholas Wyman for the past 25 years has shone...</td>\n",
       "      <td>I count myself fortunate to have spent time wi...</td>\n",
       "      <td>22</td>\n",
       "      <td>0.000757</td>\n",
       "      <td>8.7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>Nicholas Wyman</td>\n",
       "      <td>CEO IWSI Group</td>\n",
       "      <td>Nicholas Wyman for the past 25 years has shone...</td>\n",
       "      <td>Between the burgeoning unemployment rates and ...</td>\n",
       "      <td>16</td>\n",
       "      <td>0.056936</td>\n",
       "      <td>6.4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>Nicholas Wyman</td>\n",
       "      <td>CEO IWSI Group</td>\n",
       "      <td>Nicholas Wyman for the past 25 years has shone...</td>\n",
       "      <td>Congratulations to Dr. Mark Goulston on recogn...</td>\n",
       "      <td>7</td>\n",
       "      <td>0.015959</td>\n",
       "      <td>9.5</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              name        headline  \\\n",
       "1   Nicholas Wyman  CEO IWSI Group   \n",
       "2   Nicholas Wyman  CEO IWSI Group   \n",
       "3   Nicholas Wyman  CEO IWSI Group   \n",
       "5   Nicholas Wyman  CEO IWSI Group   \n",
       "10  Nicholas Wyman  CEO IWSI Group   \n",
       "\n",
       "                                                about  \\\n",
       "1   Nicholas Wyman for the past 25 years has shone...   \n",
       "2   Nicholas Wyman for the past 25 years has shone...   \n",
       "3   Nicholas Wyman for the past 25 years has shone...   \n",
       "5   Nicholas Wyman for the past 25 years has shone...   \n",
       "10  Nicholas Wyman for the past 25 years has shone...   \n",
       "\n",
       "                                              content  reactions  profanity  \\\n",
       "1   National disability advocate  Sara Hart Weir, ...         11   0.018358   \n",
       "2   Exploring in this months Talent Management & H...         44   0.008754   \n",
       "3   I count myself fortunate to have spent time wi...         22   0.000757   \n",
       "5   Between the burgeoning unemployment rates and ...         16   0.056936   \n",
       "10  Congratulations to Dr. Mark Goulston on recogn...          7   0.015959   \n",
       "\n",
       "    flesch_grade  \n",
       "1            8.8  \n",
       "2            8.8  \n",
       "3            8.7  \n",
       "5            6.4  \n",
       "10           9.5  "
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Basic ETL\n",
    "\n",
    "# Read in some additional utility functions\n",
    "\n",
    "def remove_trailing_ws(df, text_columns):\n",
    "    # For text cleaning based functions use: \n",
    "    for column in text_columns: \n",
    "        df[column] = df[column].str.strip()\n",
    "    return df\n",
    "\n",
    "def length_check_func(df,text_column, minLength,maxLength):\n",
    "    df = df[(df[text_column].str.len() > int(minLength))]\n",
    "    df = df[(df[text_column].str.len() < int(maxLength))]\n",
    "    return df\n",
    "\n",
    "def string_replace(df, text_columns, input, output):\n",
    "    # For text cleaning based functions use: \n",
    "    for column in text_columns: \n",
    "        df[column] = df[column].str.replace(input, output)\n",
    "    return df\n",
    "\n",
    "# Keep Relevant Columns\n",
    "cleaned = df[['name', 'headline', 'about', 'content', 'reactions']]\n",
    "\n",
    "# Dropping missing\n",
    "cleaned = cleaned.dropna()\n",
    "\n",
    "\n",
    "# Text Cleaning (Minimal)\n",
    "cleaned = string_replace(cleaned, ['content'], '…see more', '')\n",
    "cleaned = remove_trailing_ws(cleaned, ['content'])\n",
    "\n",
    "max_len = 10000\n",
    "min_len = 1\n",
    "\n",
    "cleaned = length_check_func(cleaned, 'content', min_len, max_len)\n",
    "\n",
    "\n",
    "# Set thresholds\n",
    "\n",
    "# Profanity Threshold\n",
    "profanity_threshold = 0.9\n",
    "\n",
    "# Target Reading Level\n",
    "min_grade_level = 2\n",
    "max_grade_level = 10\n",
    "\n",
    "# Predict Probability of Profane Language\n",
    "cleaned[\"profanity\"] = predict_prob(cleaned[\"content\"])\n",
    "\n",
    "# Filter out Profane Language\n",
    "cleaned = cleaned[cleaned[\"profanity\"] < profanity_threshold]\n",
    "cleaned = cleaned.reset_index(drop=True)\n",
    "\n",
    "# Determine Reading Level\n",
    "cleaned = check_quality(cleaned, 'content')\n",
    "\n",
    "# Filter to only those records with content between a 2nd and 10th grader\n",
    "cleaned = cleaned[(cleaned[\"flesch_grade\"] >= min_grade_level) & (cleaned[\"flesch_grade\"] <= max_grade_level)]\n",
    "cleaned.head()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "ed1f71a3",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Lost a total of 12722 records, or about 37.0%\n"
     ]
    }
   ],
   "source": [
    "num_dropped = df.shape[0] - cleaned.shape[0]\n",
    "\n",
    "print(f\"Lost a total of {num_dropped} records, or about {round(num_dropped/df.shape[0], 2) * 100}%\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "841e6ecc",
   "metadata": {},
   "source": [
    "# One Step Further\n",
    "\n",
    "If we really want to talk like an influencer, perhaps we should also additionally limit the number of records\n",
    "by the number of reactions the posts got. Let's take a look."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "fe9691a2",
   "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>name</th>\n",
       "      <th>headline</th>\n",
       "      <th>about</th>\n",
       "      <th>content</th>\n",
       "      <th>reactions</th>\n",
       "      <th>profanity</th>\n",
       "      <th>flesch_grade</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>8978</th>\n",
       "      <td>Sramana Mitra</td>\n",
       "      <td>Founder and CEO of One Million by One Million ...</td>\n",
       "      <td>Sramana Mitra is the founder and CEO of One Mi...</td>\n",
       "      <td>At what stage should a  #startup  seek pre-see...</td>\n",
       "      <td>0</td>\n",
       "      <td>0.008001</td>\n",
       "      <td>7.2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8994</th>\n",
       "      <td>Sramana Mitra</td>\n",
       "      <td>Founder and CEO of One Million by One Million ...</td>\n",
       "      <td>Sramana Mitra is the founder and CEO of One Mi...</td>\n",
       "      <td>Can you get meetings with  #VCs  and  #Angels ...</td>\n",
       "      <td>0</td>\n",
       "      <td>0.023579</td>\n",
       "      <td>8.8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>159</th>\n",
       "      <td>Karen Gross</td>\n",
       "      <td>Educator/Author (children and adult books); Se...</td>\n",
       "      <td>I'm an educator and author, writing books for ...</td>\n",
       "      <td>A piece I just wrote about a house party in Ma...</td>\n",
       "      <td>0</td>\n",
       "      <td>0.045217</td>\n",
       "      <td>4.8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>175</th>\n",
       "      <td>Karen Gross</td>\n",
       "      <td>Educator/Author (children and adult books); Se...</td>\n",
       "      <td>I'm an educator and author, writing books for ...</td>\n",
       "      <td>Consider enrolling in this course on Trauma An...</td>\n",
       "      <td>0</td>\n",
       "      <td>0.027191</td>\n",
       "      <td>7.3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>183</th>\n",
       "      <td>Karen Gross</td>\n",
       "      <td>Educator/Author (children and adult books); Se...</td>\n",
       "      <td>I'm an educator and author, writing books for ...</td>\n",
       "      <td>This suggests a new word to replace trauma in ...</td>\n",
       "      <td>0</td>\n",
       "      <td>0.003958</td>\n",
       "      <td>2.9</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>31737</th>\n",
       "      <td>Simon Sinek</td>\n",
       "      <td>Optimist and Author at Simon Sinek Inc.</td>\n",
       "      <td>Simon Sinek is an unshakable optimist who beli...</td>\n",
       "      <td>A mentor is not someone who walks ahead of us ...</td>\n",
       "      <td>103061</td>\n",
       "      <td>0.012749</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4689</th>\n",
       "      <td>Richard Branson</td>\n",
       "      <td>Founder at Virgin Group</td>\n",
       "      <td>Founder of the Virgin Group, which has gone on...</td>\n",
       "      <td>Happy 96th birthday to my biggest inspiration,...</td>\n",
       "      <td>104146</td>\n",
       "      <td>0.140939</td>\n",
       "      <td>6.3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4637</th>\n",
       "      <td>Richard Branson</td>\n",
       "      <td>Founder at Virgin Group</td>\n",
       "      <td>Founder of the Virgin Group, which has gone on...</td>\n",
       "      <td>My wonderful wife, Joan. We’ve been leaning on...</td>\n",
       "      <td>117291</td>\n",
       "      <td>0.082846</td>\n",
       "      <td>2.7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>31792</th>\n",
       "      <td>Simon Sinek</td>\n",
       "      <td>Optimist and Author at Simon Sinek Inc.</td>\n",
       "      <td>Simon Sinek is an unshakable optimist who beli...</td>\n",
       "      <td>A team is not a group of people who work toget...</td>\n",
       "      <td>118787</td>\n",
       "      <td>0.018277</td>\n",
       "      <td>2.9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8400</th>\n",
       "      <td>Ian Bremmer</td>\n",
       "      <td>President at Eurasia Group</td>\n",
       "      <td>Ian Bremmer is a political scientist who helps...</td>\n",
       "      <td>Helen is my kinda lady.</td>\n",
       "      <td>332096</td>\n",
       "      <td>0.013761</td>\n",
       "      <td>5.2</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>21290 rows × 7 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                  name                                           headline  \\\n",
       "8978     Sramana Mitra  Founder and CEO of One Million by One Million ...   \n",
       "8994     Sramana Mitra  Founder and CEO of One Million by One Million ...   \n",
       "159        Karen Gross  Educator/Author (children and adult books); Se...   \n",
       "175        Karen Gross  Educator/Author (children and adult books); Se...   \n",
       "183        Karen Gross  Educator/Author (children and adult books); Se...   \n",
       "...                ...                                                ...   \n",
       "31737      Simon Sinek            Optimist and Author at Simon Sinek Inc.   \n",
       "4689   Richard Branson                            Founder at Virgin Group   \n",
       "4637   Richard Branson                            Founder at Virgin Group   \n",
       "31792      Simon Sinek            Optimist and Author at Simon Sinek Inc.   \n",
       "8400       Ian Bremmer                         President at Eurasia Group   \n",
       "\n",
       "                                                   about  \\\n",
       "8978   Sramana Mitra is the founder and CEO of One Mi...   \n",
       "8994   Sramana Mitra is the founder and CEO of One Mi...   \n",
       "159    I'm an educator and author, writing books for ...   \n",
       "175    I'm an educator and author, writing books for ...   \n",
       "183    I'm an educator and author, writing books for ...   \n",
       "...                                                  ...   \n",
       "31737  Simon Sinek is an unshakable optimist who beli...   \n",
       "4689   Founder of the Virgin Group, which has gone on...   \n",
       "4637   Founder of the Virgin Group, which has gone on...   \n",
       "31792  Simon Sinek is an unshakable optimist who beli...   \n",
       "8400   Ian Bremmer is a political scientist who helps...   \n",
       "\n",
       "                                                 content  reactions  \\\n",
       "8978   At what stage should a  #startup  seek pre-see...          0   \n",
       "8994   Can you get meetings with  #VCs  and  #Angels ...          0   \n",
       "159    A piece I just wrote about a house party in Ma...          0   \n",
       "175    Consider enrolling in this course on Trauma An...          0   \n",
       "183    This suggests a new word to replace trauma in ...          0   \n",
       "...                                                  ...        ...   \n",
       "31737  A mentor is not someone who walks ahead of us ...     103061   \n",
       "4689   Happy 96th birthday to my biggest inspiration,...     104146   \n",
       "4637   My wonderful wife, Joan. We’ve been leaning on...     117291   \n",
       "31792  A team is not a group of people who work toget...     118787   \n",
       "8400                             Helen is my kinda lady.     332096   \n",
       "\n",
       "       profanity  flesch_grade  \n",
       "8978    0.008001           7.2  \n",
       "8994    0.023579           8.8  \n",
       "159     0.045217           4.8  \n",
       "175     0.027191           7.3  \n",
       "183     0.003958           2.9  \n",
       "...          ...           ...  \n",
       "31737   0.012749           3.0  \n",
       "4689    0.140939           6.3  \n",
       "4637    0.082846           2.7  \n",
       "31792   0.018277           2.9  \n",
       "8400    0.013761           5.2  \n",
       "\n",
       "[21290 rows x 7 columns]"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cleaned.sort_values('reactions')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "29d6b2e6",
   "metadata": {},
   "source": [
    "We can see that the number of reactions these posts from 'influencers' received appear to go \n",
    "from zero (yikes!) all the way to over 330k. Noticeably, that single post with ~330k reactions\n",
    "is simply `Helen is my kinda lady`.\n",
    "\n",
    "The fact this post received so much attention given it's lack of context is also \n",
    "a warning that there could be exogenous latent variables, such as current events, pop culture, etc, that could be driving \n",
    "the number of reactions, not necessarily the content itself. \n",
    "\n",
    "Let's be just a little more analytical:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "b1d8c2cb",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The 90th percentile is 575.0 reactions.\n"
     ]
    }
   ],
   "source": [
    "# Identify the 90th percentile of reactions over all posts\n",
    "p90 = np.quantile(cleaned.reactions, 0.9)\n",
    "\n",
    "print(f'The 90th percentile is {p90} reactions.')\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "afd0c3f2",
   "metadata": {},
   "source": [
    "For our purposes, let's focus on the top \"performing\" content to fine tune our model on"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "9c3aae04",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(2127, 7)"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cleaned = cleaned[cleaned.reactions > p90]\n",
    "cleaned.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4fbbc717",
   "metadata": {},
   "source": [
    "We're now left with 2,127 high-quality data points to experiment with \n",
    "fine tuning on. \n",
    "\n",
    "Finally, let's ask h2oGPT to provide a title for our LinkedIn Influencer content.\n",
    "This process is called `zero-shot text generation` (more on this later)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "e9d3d4b9",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loaded as API: http://h2ogpt-web.h2ogpt.svc.cluster.local/ ✔\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 5/5 [00:10<00:00,  2.05s/it]\n"
     ]
    },
    {
     "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>name</th>\n",
       "      <th>headline</th>\n",
       "      <th>about</th>\n",
       "      <th>content</th>\n",
       "      <th>reactions</th>\n",
       "      <th>profanity</th>\n",
       "      <th>flesch_grade</th>\n",
       "      <th>title</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>31918</th>\n",
       "      <td>Simon Sinek</td>\n",
       "      <td>Optimist and Author at Simon Sinek Inc.</td>\n",
       "      <td>Simon Sinek is an unshakable optimist who beli...</td>\n",
       "      <td>Bruce Deel has proven that when you take the r...</td>\n",
       "      <td>15452</td>\n",
       "      <td>0.041906</td>\n",
       "      <td>8.3</td>\n",
       "      <td>\"Empowering People Through Extreme Trust: Th...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4835</th>\n",
       "      <td>Richard Branson</td>\n",
       "      <td>Founder at Virgin Group</td>\n",
       "      <td>Founder of the Virgin Group, which has gone on...</td>\n",
       "      <td>It’s the words that count (not the spelling)  ...</td>\n",
       "      <td>3228</td>\n",
       "      <td>0.024445</td>\n",
       "      <td>2.1</td>\n",
       "      <td>\"The Power of Dyslexic Thinking: Why Words M...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>31872</th>\n",
       "      <td>Simon Sinek</td>\n",
       "      <td>Optimist and Author at Simon Sinek Inc.</td>\n",
       "      <td>Simon Sinek is an unshakable optimist who beli...</td>\n",
       "      <td>Another week of LIVE online classes is up! Bec...</td>\n",
       "      <td>1139</td>\n",
       "      <td>0.015279</td>\n",
       "      <td>9.7</td>\n",
       "      <td>\"Unlock Your Potential: Live Online Classes ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29096</th>\n",
       "      <td>Pocket Sun</td>\n",
       "      <td>Co-Founder &amp; Managing Partner at SoGal, TEDx S...</td>\n",
       "      <td>An entrepreneur, investor, and futurist, Pocke...</td>\n",
       "      <td>Can’t describe my excitement right now!! Our  ...</td>\n",
       "      <td>724</td>\n",
       "      <td>0.010127</td>\n",
       "      <td>9.2</td>\n",
       "      <td>\"SoGal Ventures' Seed Investment in Everlywe...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5092</th>\n",
       "      <td>Richard Branson</td>\n",
       "      <td>Founder at Virgin Group</td>\n",
       "      <td>Founder of the Virgin Group, which has gone on...</td>\n",
       "      <td>Waving goodbye to the  Virgin Trains  West Coa...</td>\n",
       "      <td>796</td>\n",
       "      <td>0.058986</td>\n",
       "      <td>6.0</td>\n",
       "      <td>\"Virgin Trains Bids Farewell to West Coast M...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                  name                                           headline  \\\n",
       "31918      Simon Sinek            Optimist and Author at Simon Sinek Inc.   \n",
       "4835   Richard Branson                            Founder at Virgin Group   \n",
       "31872      Simon Sinek            Optimist and Author at Simon Sinek Inc.   \n",
       "29096       Pocket Sun  Co-Founder & Managing Partner at SoGal, TEDx S...   \n",
       "5092   Richard Branson                            Founder at Virgin Group   \n",
       "\n",
       "                                                   about  \\\n",
       "31918  Simon Sinek is an unshakable optimist who beli...   \n",
       "4835   Founder of the Virgin Group, which has gone on...   \n",
       "31872  Simon Sinek is an unshakable optimist who beli...   \n",
       "29096  An entrepreneur, investor, and futurist, Pocke...   \n",
       "5092   Founder of the Virgin Group, which has gone on...   \n",
       "\n",
       "                                                 content  reactions  \\\n",
       "31918  Bruce Deel has proven that when you take the r...      15452   \n",
       "4835   It’s the words that count (not the spelling)  ...       3228   \n",
       "31872  Another week of LIVE online classes is up! Bec...       1139   \n",
       "29096  Can’t describe my excitement right now!! Our  ...        724   \n",
       "5092   Waving goodbye to the  Virgin Trains  West Coa...        796   \n",
       "\n",
       "       profanity  flesch_grade  \\\n",
       "31918   0.041906           8.3   \n",
       "4835    0.024445           2.1   \n",
       "31872   0.015279           9.7   \n",
       "29096   0.010127           9.2   \n",
       "5092    0.058986           6.0   \n",
       "\n",
       "                                                   title  \n",
       "31918    \"Empowering People Through Extreme Trust: Th...  \n",
       "4835     \"The Power of Dyslexic Thinking: Why Words M...  \n",
       "31872    \"Unlock Your Potential: Live Online Classes ...  \n",
       "29096    \"SoGal Ventures' Seed Investment in Everlywe...  \n",
       "5092     \"Virgin Trains Bids Farewell to West Coast M...  "
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sample_df = cleaned.sample(5)\n",
    "\n",
    "from gradio_client import Client\n",
    "import ast\n",
    "from pprint import pprint\n",
    "\n",
    "# Dedicated h2oGPT\n",
    "HOST_URL = \"http://h2ogpt-web.h2ogpt.svc.cluster.local\"\n",
    "H2OGPT_KEY = \"f74f043e-45fc-4dfe-9c33-55a4720427f6\"\n",
    "    \n",
    "client = Client(HOST_URL)\n",
    "\n",
    "from tqdm import tqdm\n",
    "tqdm.pandas()\n",
    "\n",
    "def generate_title(content):\n",
    "    \n",
    "    #try:\n",
    "    summarize_prompt = 'You are a helpful, respectful and honest assistant the specializes in generating accurate titles of LinkedIn posts. Provide a title for the following post. The title should be a single sentence, not using bullet points. Only include the title in the response. The LinkedIn post is: ' + content\n",
    "    kwargs = dict(\n",
    "        instruction_nochat=summarize_prompt, \n",
    "        h2ogpt_key = H2OGPT_KEY)\n",
    "    \n",
    "    response = client.predict(str(dict(kwargs)), api_name='/submit_nochat_api')\n",
    "    reply = ast.literal_eval(response)['response']\n",
    "    #except:\n",
    "        #reply = 'NA'\n",
    "    return reply\n",
    "        \n",
    "sample_df['title'] = sample_df.progress_apply(lambda row :   generate_title(row['content']), axis=1)\n",
    "sample_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "f39448c6",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CONTENT:\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "['Bruce Deel has proven that when you take the risk to trust people, they rise to the occasion. That\\'s why he practices \"extreme trust\"\\u200b through his organization City of Refuge, and why I am so excited to publish his book \"Trust First\"\\u200b with #OptimismPress.',\n",
       " 'It’s the words that count (not the spelling)  https://virg.in/58P   Made By Dyslexia',\n",
       " 'Another week of LIVE online classes is up! Because they are LIVE, we offer classes at various times to help reach as many of you as we can. To sign up, visit  simonsinek.com/shop . Here’s what’s happening this week: MONDAY How to Lead Through Uncertainty @ 10am ET TUESDAY How to Pivot @ 7pm ET WEDNESDAY How to Build a Strong Culture @ 10am ET THURSDAY How to Deliver and Accept Feedback Effectively @ 10am ET Jumpstart Your WHY: Office Hours @ 3pm ET FRIDAY The Infinite Game: How to Advance Your Just Cause @ 9pm ET',\n",
       " 'Can’t describe my excitement right now!! Our  SoGal Ventures  seed stage investment in March 2017 is now a $1.3 Billion dollar unicorn 🦄 !!! BlackRock and many amazing new investors including women focused  Portfolia  joined this $175M Series D round.  We are incredibly fortunate to be supporters of  Julia Cheek  and the diverse  Everlywell  team!! They’ve been surpassing expectations every month, every quarter, and every year. They showed amazing leadership this year during Covid and pioneered at home Covid tests despite being a 90-person startup. As a solo founder, Julia not only led the company to be successful beyond metrics, but also created an inclusive space for her team to grow, scale, and thrive.   3 years ago when we started the fund, barely anyone believed  Elizabeth Galbut  and I could be VCs, or that focusing on investing in women makes any sense. Yet here we are!! Standards are meant to be broken and broadened. I hope this story inspires you to break into anything you want to be in, and use your difference to make a real, big impact 👊🏼 #startup   #venturecapital   #investment   #digitalhealth   #entrepreneur',\n",
       " 'Waving goodbye to the  Virgin Trains  West Coast Mainline  https://virg.in/iuk']"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(f'CONTENT:')\n",
    "sample_df['content'].to_list()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "af29b9b2",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "TITLE:\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "['  \"Empowering People Through Extreme Trust: The City of Refuge Story and the Power of Trust First\"',\n",
       " '  \"The Power of Dyslexic Thinking: Why Words Matter More Than Spelling\"',\n",
       " '  \"Unlock Your Potential: Live Online Classes to Help You Thrive in Uncertainty\"',\n",
       " '  \"SoGal Ventures\\' Seed Investment in Everlywell Now Valued at $1.3 Billion: A Journey to Breaking Standards and Making an Impact\"',\n",
       " '  \"Virgin Trains Bids Farewell to West Coast Mainline: A New Era Begins\"']"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(f'TITLE:')\n",
    "sample_df['title'].to_list()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "10fbf629",
   "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>name</th>\n",
       "      <th>headline</th>\n",
       "      <th>about</th>\n",
       "      <th>content</th>\n",
       "      <th>reactions</th>\n",
       "      <th>profanity</th>\n",
       "      <th>flesch_grade</th>\n",
       "      <th>instruction</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>5247</th>\n",
       "      <td>Richard Branson</td>\n",
       "      <td>Founder at Virgin Group</td>\n",
       "      <td>Founder of the Virgin Group, which has gone on...</td>\n",
       "      <td>Thrilled to celebrate Virgin Atlantic ’s new r...</td>\n",
       "      <td>4157</td>\n",
       "      <td>0.160989</td>\n",
       "      <td>6.8</td>\n",
       "      <td>Write a LinkedIn post in the style of an influ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>31937</th>\n",
       "      <td>Simon Sinek</td>\n",
       "      <td>Optimist and Author at Simon Sinek Inc.</td>\n",
       "      <td>Simon Sinek is an unshakable optimist who beli...</td>\n",
       "      <td>Great leaders don't see people as a commodity ...</td>\n",
       "      <td>15282</td>\n",
       "      <td>0.033452</td>\n",
       "      <td>9.7</td>\n",
       "      <td>Write a LinkedIn post in the style of an influ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27972</th>\n",
       "      <td>Vani Kola</td>\n",
       "      <td>Managing Director at Kalaari</td>\n",
       "      <td>Vani Kola is the Founder and Managing Director...</td>\n",
       "      <td>Going to the family jeweller was the custom in...</td>\n",
       "      <td>583</td>\n",
       "      <td>0.021146</td>\n",
       "      <td>7.9</td>\n",
       "      <td>Write a LinkedIn post in the style of an influ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6255</th>\n",
       "      <td>Mohamed El-Erian</td>\n",
       "      <td>President at Queens'​ College, Cambridge</td>\n",
       "      <td>Dr Mohamed A. El-Erian is President of Queens'...</td>\n",
       "      <td>Happy Father's Day to all the dads out there. ...</td>\n",
       "      <td>3757</td>\n",
       "      <td>0.022549</td>\n",
       "      <td>6.1</td>\n",
       "      <td>Write a LinkedIn post in the style of an influ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>31955</th>\n",
       "      <td>Simon Sinek</td>\n",
       "      <td>Optimist and Author at Simon Sinek Inc.</td>\n",
       "      <td>Simon Sinek is an unshakable optimist who beli...</td>\n",
       "      <td>The greatest contribution of a leader is to ma...</td>\n",
       "      <td>71183</td>\n",
       "      <td>0.023041</td>\n",
       "      <td>7.6</td>\n",
       "      <td>Write a LinkedIn post in the style of an influ...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                   name                                  headline  \\\n",
       "5247    Richard Branson                   Founder at Virgin Group   \n",
       "31937       Simon Sinek   Optimist and Author at Simon Sinek Inc.   \n",
       "27972         Vani Kola              Managing Director at Kalaari   \n",
       "6255   Mohamed El-Erian  President at Queens'​ College, Cambridge   \n",
       "31955       Simon Sinek   Optimist and Author at Simon Sinek Inc.   \n",
       "\n",
       "                                                   about  \\\n",
       "5247   Founder of the Virgin Group, which has gone on...   \n",
       "31937  Simon Sinek is an unshakable optimist who beli...   \n",
       "27972  Vani Kola is the Founder and Managing Director...   \n",
       "6255   Dr Mohamed A. El-Erian is President of Queens'...   \n",
       "31955  Simon Sinek is an unshakable optimist who beli...   \n",
       "\n",
       "                                                 content  reactions  \\\n",
       "5247   Thrilled to celebrate Virgin Atlantic ’s new r...       4157   \n",
       "31937  Great leaders don't see people as a commodity ...      15282   \n",
       "27972  Going to the family jeweller was the custom in...        583   \n",
       "6255   Happy Father's Day to all the dads out there. ...       3757   \n",
       "31955  The greatest contribution of a leader is to ma...      71183   \n",
       "\n",
       "       profanity  flesch_grade  \\\n",
       "5247    0.160989           6.8   \n",
       "31937   0.033452           9.7   \n",
       "27972   0.021146           7.9   \n",
       "6255    0.022549           6.1   \n",
       "31955   0.023041           7.6   \n",
       "\n",
       "                                             instruction  \n",
       "5247   Write a LinkedIn post in the style of an influ...  \n",
       "31937  Write a LinkedIn post in the style of an influ...  \n",
       "27972  Write a LinkedIn post in the style of an influ...  \n",
       "6255   Write a LinkedIn post in the style of an influ...  \n",
       "31955  Write a LinkedIn post in the style of an influ...  "
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Optional alternative for fine tuning - create an instruction.\n",
    "cleaned['instruction'] = ('Write a LinkedIn post in the style of an influencer whom has the title of '\n",
    "  + cleaned['headline'] + ' and can be described by the following: ' \n",
    "  + cleaned['about'])\n",
    "\n",
    "cleaned.sample(5)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b000f0d7",
   "metadata": {},
   "source": [
    "# Output Dataset\n",
    "\n",
    "Now we're ready to store our data set out and experiment with fine-tuning\n",
    "in Lab # 2 with H2O LLM Studio. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "7ba4567e",
   "metadata": {},
   "outputs": [],
   "source": [
    "#######################################################################################\n",
    "# WARNING! This could take a very long time using the public facing h2oGPT endpoint. ##\n",
    "#######################################################################################\n",
    "\n",
    "# Apply to full data set\n",
    "# cleaned['title'] = cleaned.progress_apply(lambda row :   generate_title(row['content']), axis=1)\n",
    "\n",
    "# Output locally\n",
    "# cleaned.to_csv('influencers_data_prepared.csv', index=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "43da402b",
   "metadata": {},
   "source": [
    "# 🎉 **CONGRATULATIONS!** You have completed this lab!\n",
    "---"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7e6dd72c",
   "metadata": {},
   "source": [
    "# 📚 **EXTENDED TASKS** \n",
    "\n",
    "# Understanding Enterprise Grade Dataprep with H2O LLM Data Studio\n",
    "\n",
    "\n",
    "We've done an OK job manually looking through this data set in anticipation of \n",
    "fine-tuning a model on it. Specifically, we cleaned the text data, filtered it, \n",
    "and asked generative AI (h2oGPT) to create a title for each post. \n",
    "\n",
    "In practice, however, this won't scale very well. H2O.ai as created a tool called\n",
    "`H2O LLM DataStudio` that enables you to run far more thorough data engineering \n",
    "pipelines for textual data. \n",
    "\n",
    "As the screenshot below shows, users can prepare their text data for numerous language \n",
    "modeling tasks. \n",
    "\n",
    "These include:\n",
    "\n",
    "- Data Augmentation\n",
    "- Text Cleaning\n",
    "- Profanity Check \n",
    "- Text Quality Check\n",
    "- Length Checker \n",
    "- Valid Question\n",
    "- Pad Sequence\n",
    "- Truncate Sequence by Score \n",
    "- Compression Ratio filter\n",
    "- Boundary Marking\n",
    "- Sensitive Info Checker\n",
    "- RLHF Protection\n",
    "- Language Understanding \n",
    "- Data Deduplication\n",
    "- Toxicity Detection\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "---\n",
    "<img src=\"static/h2o-data-studio.png\" alt=\"drawing\" width=\"800\"/> \n",
    "---\n",
    "\n",
    "\n",
    "---"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9cce20fb",
   "metadata": {},
   "source": [
    "# 🐍 Python CLI\n",
    "\n",
    "Not only is LLM Data Studio a standalone Wave application and UI, it also \n",
    "supports programmatic data engineering pipelines. The following code snippet\n",
    "provides an example of configuring a data engineering pipeline in Python.\n",
    "\n",
    "For a deeper dive check out the blog post:\n",
    "https://h2o.ai/blog/2023/streamlining-data-preparation-for-fine-tuning-of-large-language-models/"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ba89253a",
   "metadata": {},
   "source": [
    "```python\n",
    "# Importing required modules\n",
    "from datastudio import prep\n",
    "import pandas as pd\n",
    "\n",
    "# Initializing configuration parameters for Question Answer Workflow\n",
    "cfg = {\n",
    "    \"type\": \"qa\",\n",
    "    'config': { \n",
    "        'augmentation': [],\n",
    "        \"text_clean\": {\n",
    "            \"cols\": [],\n",
    "            \"funcs\": [\"new_line\",\"whitespace\",\"lower_case\",\"urls\",\"html\",\"ascii\"]\n",
    "        },\n",
    "        'length_checker': {\n",
    "            'max_answer_length': 10000,\n",
    "            'max_context_length': 10000,\n",
    "            'max_question_length': 10000,\n",
    "            'min_answer_length': 0,\n",
    "            'min_context_length': 0,\n",
    "            'min_question_length': 0\n",
    "        },\n",
    "        'padding': {'max_length': '600'},\n",
    "        'profanity_checker': 0.3,\n",
    "        'quality_checker': {'max_grade': 30, 'min_grade': 15},\n",
    "        'relevance_checker': 0.15,\n",
    "        'truncate': {'max_length': '10000', 'ratio': 0.15}\n",
    "    },\n",
    "    'dataflow': ['augmentation', 'text_cleaner', 'profanity_checker', 'length_checker', \n",
    "                 'grade_checker', 'qna_relevance', 'padding_sequence', 'truncate_sequence'],\n",
    "    'datasets': [{\n",
    "        'cols': {'answer': 'answer', 'context': 'context', 'question': 'question'},\n",
    "        'path': ['input/qna.csv']\n",
    "    }],\n",
    "    'is_hum_bot': 0,\n",
    "    \"selected_aug_dataset\": 0\n",
    "}\n",
    "\n",
    "# Executing data preparation pipeline\n",
    "output_df = prep.prepare_dataset(cfg)\n",
    "\n",
    "# Displaying output DataFrame\n",
    "output_df.head()\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d984c7f6",
   "metadata": {},
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python",
   "language": "python",
   "name": "python_kubernetes"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}