1. EDA , Wrangling, and Initial Preprocessing.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# EDA, Wrangling, and Initial Preprocessing"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "To start, I was choosing between these datasets:\n",
    "* Customer Support on Twitter https://www.kaggle.com/thoughtvector/customer-support-on-twitter\n",
    "* Ubuntu Dialogue Corpus https://www.kaggle.com/rtatman/ubuntu-dialogue-corpus\n",
    "* Relational Services in Customer Service (RSICS) https://nextit-public.s3-us-west-2.amazonaws.com/rsics.html"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Since I am making a chatbot for helping with application problems, the following block defines the format of intents I want to classify as the keys. The values are the predefined responses that I might want to put. This is just a toy example to illustrate this concept for context. The unsupervised learning happens in the following notebook, this notebook is just to apply my wrangling and entire preprocessing pipeline and familiarize myself with my data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Versions\n",
      "Pandas: 1.0.5\n",
      "Numpy: 1.18.5\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/anaconda3/envs/chatbot/lib/python3.7/site-packages/sklearn/utils/deprecation.py:143: FutureWarning: The sklearn.feature_extraction.stop_words module is  deprecated in version 0.22 and will be removed in version 0.24. The corresponding classes / functions should instead be imported from sklearn.feature_extraction.text. Anything that cannot be imported from sklearn.feature_extraction.text is now part of the private API.\n",
      "  warnings.warn(message, FutureWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "re: 2.2.1\n",
      "spaCy: 2.3.0\n",
      "nltk: 3.5\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/anaconda3/envs/chatbot/lib/python3.7/site-packages/ipykernel_launcher.py:50: TqdmDeprecationWarning: This function will be removed in tqdm==5.0.0\n",
      "Please use `tqdm.notebook.tqdm` instead of `tqdm.tqdm_notebook`\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "57629fee320c4929a4ad7acff8f8d473",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=1.0, bar_style='info', max=1.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/anaconda3/envs/chatbot/lib/python3.7/site-packages/tqdm/std.py:668: FutureWarning: The Panel class is removed from pandas. Accessing it from the top-level namespace will also be removed in the next version\n",
      "  from pandas import Panel\n"
     ]
    }
   ],
   "source": [
    "print(\"Versions\")\n",
    "# Data science\n",
    "import pandas as pd\n",
    "print(f\"Pandas: {pd.__version__}\")\n",
    "import numpy as np\n",
    "print(f\"Numpy: {np.__version__}\")\n",
    "\n",
    "# Visualization \n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "sns.set(style=\"ticks\", color_codes=True)\n",
    "from bokeh.io import output_file, show\n",
    "from bokeh.plotting import figure\n",
    "from bokeh.models import ColorBar, ColumnDataSource\n",
    "from bokeh.colors import Color\n",
    "\n",
    "# Making my visualizations pretty\n",
    "sns.set_style('whitegrid')\n",
    "\n",
    "# Cool text visualization\n",
    "import scattertext as st\n",
    "\n",
    "# Regular Expressions\n",
    "import re\n",
    "print(f're: {re.__version__}')\n",
    "\n",
    "# Preprocessing\n",
    "import spacy\n",
    "print(f'spaCy: {spacy.__version__}')\n",
    "import nltk\n",
    "print(f'nltk: {nltk.__version__}')\n",
    "from spellchecker import SpellChecker\n",
    "from nltk.corpus import stopwords\n",
    "from nltk.tokenize import word_tokenize \n",
    "from sklearn.feature_extraction.text import CountVectorizer\n",
    "from nltk.tokenize import casual_tokenize\n",
    "from nltk.tokenize import TweetTokenizer\n",
    "from nltk.stem import WordNetLemmatizer\n",
    "from nltk.corpus import wordnet\n",
    "import yaml\n",
    "\n",
    "# Emoji analysis and language detection\n",
    "from spacymoji import Emoji\n",
    "import spacy_cld\n",
    "# For the counter\n",
    "from collections import Counter\n",
    "\n",
    "# Cool progress bars\n",
    "from tqdm import tqdm_notebook as tqdm\n",
    "tqdm().pandas()  # Enable tracking of execution progress"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Twitter\n",
    "I ended up settling with the Twitter data (see comparison at the back of this notebook if you're interested in why) in order to get my labelled data.\n",
    "\n",
    "Notes about dataset:\n",
    "* It's from 2017 to 2018\n",
    "* From another Kernel I saw that AmazonSupport contained the most tweets, and amazingly, it turns out, not all of the tweets are in English!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1800x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 11.4 s, sys: 1.9 s, total: 13.3 s\n",
      "Wall time: 14.7 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "# Reading in all the Twitter data\n",
    "all_data = pd.read_csv('data/twitter/twcs.csv')\n",
    "\n",
    "# Finding the distribution of all authors\n",
    "count = all_data.groupby(\"author_id\")[\"text\"].count()\n",
    "\n",
    "# Showing only the authors that appear more than 15000 times\n",
    "c = count[count>15000].plot(kind='bar',figsize=(15, 5), color='#00acee')\n",
    "c.set_xlabel('')\n",
    "c.set_ylabel('Frequency')\n",
    "plt.title('Most of the Data is AmazonHelp')\n",
    "\n",
    "# Saving my plot\n",
    "plt.savefig('visualizations/authors.png')\n",
    "\n",
    "plt.show()\n",
    "\n",
    "# There are 108 unique authors that are companies and 702669 customers\n",
    "authors_count = (len(all_data[all_data.inbound == False].author_id.unique()), len(all_data[all_data.inbound == True].author_id.unique()))\n",
    "\n",
    "# Visualizing it\n",
    "plt.figure(figsize =(25,5))\n",
    "x = ['Unique Companies','Unique Customers']\n",
    "ax = plt.barh(x, authors_count, color = '#00acee')\n",
    "plt.title('There are way more unique customers than unique companies')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Data Dictionary:\n",
    "* Tweet_id: The unique ID for this tweet\n",
    "* Author_id: The unique ID for this tweet author (anonymized for non-company users)\n",
    "* Inbound: Whether or not the tweet was sent (inbound) to a company\n",
    "* Created_at: When the tweet was created\n",
    "* Text: The text content of the tweet\n",
    "* Response_tweet_id: The tweet that responded to this one, if any\n",
    "* In_response_to_tweet_id: The tweet this tweet was in response to, if any"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "All data has shape (2811774, 7)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<style  type=\"text/css\" >\n",
       "    #T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow0_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  cyan;\n",
       "            background-color:  #08306b;\n",
       "            color:  #f1f1f1;\n",
       "        }    #T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow0_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow0_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  cyan;\n",
       "        }    #T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow0_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow0_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow0_col5 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  cyan;\n",
       "        }    #T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow0_col6 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  cyan;\n",
       "            background-color:  #08306b;\n",
       "            color:  #f1f1f1;\n",
       "        }    #T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow1_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  cyan;\n",
       "            background-color:  #08306b;\n",
       "            color:  #f1f1f1;\n",
       "        }    #T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow1_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow1_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  cyan;\n",
       "        }    #T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow1_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow1_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow1_col5 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow1_col6 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  cyan;\n",
       "            background-color:  #000000;\n",
       "            color:  #f1f1f1;\n",
       "        }    #T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow2_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  cyan;\n",
       "            background-color:  #f7fbff;\n",
       "            color:  #000000;\n",
       "        }    #T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow2_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow2_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  cyan;\n",
       "        }    #T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow2_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow2_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow2_col5 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  cyan;\n",
       "        }    #T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow2_col6 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  cyan;\n",
       "            background-color:  #f7fbff;\n",
       "            color:  #000000;\n",
       "        }    #T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow3_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  cyan;\n",
       "            background-color:  #08306b;\n",
       "            color:  #f1f1f1;\n",
       "        }    #T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow3_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow3_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  cyan;\n",
       "        }    #T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow3_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow3_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow3_col5 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  cyan;\n",
       "        }    #T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow3_col6 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  cyan;\n",
       "            background-color:  #08306b;\n",
       "            color:  #f1f1f1;\n",
       "        }    #T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow4_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  cyan;\n",
       "            background-color:  #08306b;\n",
       "            color:  #f1f1f1;\n",
       "        }    #T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow4_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow4_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  cyan;\n",
       "        }    #T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow4_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow4_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow4_col5 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow4_col6 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  cyan;\n",
       "            background-color:  #000000;\n",
       "            color:  #f1f1f1;\n",
       "        }</style><table id=\"T_1161c73e_d2b4_11ea_aff7_acbc32ce112f\" ><thead>    <tr>        <th class=\"blank level0\" ></th>        <th class=\"col_heading level0 col0\" >tweet_id</th>        <th class=\"col_heading level0 col1\" >author_id</th>        <th class=\"col_heading level0 col2\" >inbound</th>        <th class=\"col_heading level0 col3\" >created_at</th>        <th class=\"col_heading level0 col4\" >text</th>        <th class=\"col_heading level0 col5\" >response_tweet_id</th>        <th class=\"col_heading level0 col6\" >in_response_to_tweet_id</th>    </tr></thead><tbody>\n",
       "                <tr>\n",
       "                        <th id=\"T_1161c73e_d2b4_11ea_aff7_acbc32ce112flevel0_row0\" class=\"row_heading level0 row0\" >2811769</th>\n",
       "                        <td id=\"T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow0_col0\" class=\"data row0 col0\" >2987947</td>\n",
       "                        <td id=\"T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow0_col1\" class=\"data row0 col1\" >sprintcare</td>\n",
       "                        <td id=\"T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow0_col2\" class=\"data row0 col2\" >False</td>\n",
       "                        <td id=\"T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow0_col3\" class=\"data row0 col3\" >Wed Nov 22 08:43:51 +0000 2017</td>\n",
       "                        <td id=\"T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow0_col4\" class=\"data row0 col4\" >@823869 Hey, we'd be happy to look into this for you. Please send us a Direct Message, so that we can further assist you. -TN</td>\n",
       "                        <td id=\"T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow0_col5\" class=\"data row0 col5\" >nan</td>\n",
       "                        <td id=\"T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow0_col6\" class=\"data row0 col6\" >2987948.000000</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_1161c73e_d2b4_11ea_aff7_acbc32ce112flevel0_row1\" class=\"row_heading level0 row1\" >2811770</th>\n",
       "                        <td id=\"T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow1_col0\" class=\"data row1 col0\" >2987948</td>\n",
       "                        <td id=\"T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow1_col1\" class=\"data row1 col1\" >823869</td>\n",
       "                        <td id=\"T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow1_col2\" class=\"data row1 col2\" >True</td>\n",
       "                        <td id=\"T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow1_col3\" class=\"data row1 col3\" >Wed Nov 22 08:35:16 +0000 2017</td>\n",
       "                        <td id=\"T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow1_col4\" class=\"data row1 col4\" >@115714 wtf!? I’ve been having really shitty service all day !! Get your shit together!</td>\n",
       "                        <td id=\"T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow1_col5\" class=\"data row1 col5\" >2987947</td>\n",
       "                        <td id=\"T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow1_col6\" class=\"data row1 col6\" >nan</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_1161c73e_d2b4_11ea_aff7_acbc32ce112flevel0_row2\" class=\"row_heading level0 row2\" >2811771</th>\n",
       "                        <td id=\"T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow2_col0\" class=\"data row2 col0\" >2812240</td>\n",
       "                        <td id=\"T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow2_col1\" class=\"data row2 col1\" >121673</td>\n",
       "                        <td id=\"T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow2_col2\" class=\"data row2 col2\" >True</td>\n",
       "                        <td id=\"T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow2_col3\" class=\"data row2 col3\" >Thu Nov 23 04:13:07 +0000 2017</td>\n",
       "                        <td id=\"T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow2_col4\" class=\"data row2 col4\" >@143549 @sprintcare You have to go to https://t.co/v2tMHeTL7Q and ask them to add the Hulu service to your account. After you get the text, you’re good to go to https://t.co/gFyUq1kJTK</td>\n",
       "                        <td id=\"T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow2_col5\" class=\"data row2 col5\" >nan</td>\n",
       "                        <td id=\"T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow2_col6\" class=\"data row2 col6\" >2812239.000000</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_1161c73e_d2b4_11ea_aff7_acbc32ce112flevel0_row3\" class=\"row_heading level0 row3\" >2811772</th>\n",
       "                        <td id=\"T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow3_col0\" class=\"data row3 col0\" >2987949</td>\n",
       "                        <td id=\"T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow3_col1\" class=\"data row3 col1\" >AldiUK</td>\n",
       "                        <td id=\"T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow3_col2\" class=\"data row3 col2\" >False</td>\n",
       "                        <td id=\"T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow3_col3\" class=\"data row3 col3\" >Wed Nov 22 08:31:24 +0000 2017</td>\n",
       "                        <td id=\"T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow3_col4\" class=\"data row3 col4\" >@823870 Sounds delicious, Sarah! 😋 https://t.co/7uQPwyH1B6</td>\n",
       "                        <td id=\"T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow3_col5\" class=\"data row3 col5\" >nan</td>\n",
       "                        <td id=\"T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow3_col6\" class=\"data row3 col6\" >2987950.000000</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_1161c73e_d2b4_11ea_aff7_acbc32ce112flevel0_row4\" class=\"row_heading level0 row4\" >2811773</th>\n",
       "                        <td id=\"T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow4_col0\" class=\"data row4 col0\" >2987950</td>\n",
       "                        <td id=\"T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow4_col1\" class=\"data row4 col1\" >823870</td>\n",
       "                        <td id=\"T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow4_col2\" class=\"data row4 col2\" >True</td>\n",
       "                        <td id=\"T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow4_col3\" class=\"data row4 col3\" >Tue Nov 21 22:01:04 +0000 2017</td>\n",
       "                        <td id=\"T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow4_col4\" class=\"data row4 col4\" >@AldiUK  warm sloe gin mince pies with ice cream - the best thing ever! #notjustxmas#allyearround</td>\n",
       "                        <td id=\"T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow4_col5\" class=\"data row4 col5\" >2987951,2987949</td>\n",
       "                        <td id=\"T_1161c73e_d2b4_11ea_aff7_acbc32ce112frow4_col6\" class=\"data row4 col6\" >nan</td>\n",
       "            </tr>\n",
       "    </tbody></table>"
      ],
      "text/plain": [
       "<pandas.io.formats.style.Styler at 0x1928e31950>"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# All data shape\n",
    "print('All data has shape {}'.format(all_data.shape))\n",
    "\n",
    "# Converting columns to the right datatypes\n",
    "all_data['text'] = all_data['text'].astype('string')\n",
    "all_data['author_id'] = all_data['author_id'].astype('string')\n",
    "\n",
    "# This function utilizes Panda's Style to helps me encode data type and \n",
    "# magnitude information to my dataframe view\n",
    "show = lambda x: x.tail(5).style.set_properties(**{'background-color': 'black',                                                   \n",
    "                                    'color': 'lawngreen',                       \n",
    "                                    'border-color': 'white'})\\\n",
    ".applymap(lambda x: f\"color: {'lawngreen' if isinstance(x,str) else 'cyan'}\")\\\n",
    ".background_gradient(cmap='Blues')\n",
    "\n",
    "show(all_data)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "You can see some data types are off, like the float."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Data Wrangling\n",
    "\n",
    "**Requirements**:\n",
    "* **Inbound**: Since the first step I want to achieve is intent classification and to figure out what the customer is saying to the company, I only want data that goes from the customer to the company (inbound data). \n",
    "* **Just Apple data:** My first modeling step is to do some form of clustering to classify the intents for each of these tweets, doing so might be easier and more fruitful if I narrow the data to just the apple data because they will be in a somewhat similar domain. I can expand this over to other datasets.\n",
    "* **First Tweet**\n",
    "\n",
    "Later, I would like my chatbot to be able to respond to make *follow up* responses to my customers."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(106860, 7)"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "apple_data = all_data[all_data['author_id'] == 'AppleSupport']\n",
    "apple_data.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "There are 106860 rows of Apple data, which is 3.8 percent of the entire data. These are all data from Apple to a customer, I want the data from the customer to Apple for each of these Apple customer service responses, and this can be found in the reponse_tweet_id.\n",
    "\n",
    "Now I drop the rows with a NaN value for the in response to tweet id column, meaning that the original customer inquiry was not available for that interaction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(False    106719\n",
       " Name: inbound, dtype: int64,\n",
       " (106719, 7))"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "apple_data.dropna(subset = ['in_response_to_tweet_id'],inplace = True)\n",
    "apple_data['inbound'].value_counts(), apple_data.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Nice, it seems like we still have plenty of data to work with (106719 rows).\n",
    "\n",
    "All of this is inbound data, because it's authored by Apple. Now I want to find all the original tweets from the customer to Apple Support and return them as a text in the Dataframe as opposed to an ID.\n",
    "\n",
    "In other words, I change ```in_response_to_tweet_id``` from it's ID to it's text so we can see the what customer question this Apple Support Tweet was responding to."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                                        inbound_text  \\\n",
      "0  @AppleSupport The newest update. I️ made sure ...   \n",
      "1             @AppleSupport  https://t.co/NV0yucs0lB   \n",
      "2  @AppleSupport Tried resetting my settings .. r...   \n",
      "3  @AppleSupport This is what it looks like https...   \n",
      "4  @AppleSupport I️ have an iPhone 7 Plus and yes...   \n",
      "\n",
      "                                       outbound_text  \n",
      "0  @115854 Lets take a closer look into this issu...  \n",
      "1  @115854 We're here for you. Which version of t...  \n",
      "2  @115855 Let's go to DM for the next steps. DM ...  \n",
      "3  @115855 Any steps tried since it started last ...  \n",
      "4  @115855 That's great it has iOS 11.1 as we can...  \n"
     ]
    }
   ],
   "source": [
    "# Converting the IDs to integers to match the tweet_id column for joining\n",
    "apple_data.in_response_to_tweet_id = apple_data.in_response_to_tweet_id.astype('int64')\n",
    "\n",
    "# Doing an inner join to get the respective tweets back from the all_data from and get the response texts\n",
    "apple_data = pd.merge(all_data[['tweet_id','text']],apple_data,left_on='tweet_id', right_on='in_response_to_tweet_id')\n",
    "\n",
    "# Dropping the columns that doesn't encode useful information and renaming the columns nicely for intepretability\n",
    "apple_data.drop(['tweet_id_x','tweet_id_y','in_response_to_tweet_id','inbound'], axis = 1, inplace = True)\n",
    "\n",
    "# Renaming the columns for intepretability\n",
    "apple_data.rename(columns = {'text_x':'inbound_text','text_y':'outbound_text'},inplace = True)\n",
    "\n",
    "# Subsetting to inbound and outbound data\n",
    "text = apple_data[['inbound_text','outbound_text']]\n",
    "\n",
    "# Previewing\n",
    "print(text.head())\n",
    "\n",
    "# Saving the real, raw inbound data\n",
    "text.to_pickle('objects/raw_text.pkl')\n",
    "\n",
    "inbound = text['inbound_text']\n",
    "outbound = text['outbound_text']\n",
    "\n",
    "# Unpacking helps me expand my data at each row for inspection\n",
    "# [*text['outbound']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style  type=\"text/css\" >\n",
       "    #T_17c26732_d2b4_11ea_aff7_acbc32ce112frow0_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_17c26732_d2b4_11ea_aff7_acbc32ce112frow0_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_17c26732_d2b4_11ea_aff7_acbc32ce112frow0_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_17c26732_d2b4_11ea_aff7_acbc32ce112frow0_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_17c26732_d2b4_11ea_aff7_acbc32ce112frow0_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_17c26732_d2b4_11ea_aff7_acbc32ce112frow1_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_17c26732_d2b4_11ea_aff7_acbc32ce112frow1_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_17c26732_d2b4_11ea_aff7_acbc32ce112frow1_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_17c26732_d2b4_11ea_aff7_acbc32ce112frow1_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_17c26732_d2b4_11ea_aff7_acbc32ce112frow1_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  cyan;\n",
       "        }    #T_17c26732_d2b4_11ea_aff7_acbc32ce112frow2_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_17c26732_d2b4_11ea_aff7_acbc32ce112frow2_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_17c26732_d2b4_11ea_aff7_acbc32ce112frow2_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_17c26732_d2b4_11ea_aff7_acbc32ce112frow2_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_17c26732_d2b4_11ea_aff7_acbc32ce112frow2_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  cyan;\n",
       "        }    #T_17c26732_d2b4_11ea_aff7_acbc32ce112frow3_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_17c26732_d2b4_11ea_aff7_acbc32ce112frow3_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_17c26732_d2b4_11ea_aff7_acbc32ce112frow3_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_17c26732_d2b4_11ea_aff7_acbc32ce112frow3_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_17c26732_d2b4_11ea_aff7_acbc32ce112frow3_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  cyan;\n",
       "        }    #T_17c26732_d2b4_11ea_aff7_acbc32ce112frow4_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_17c26732_d2b4_11ea_aff7_acbc32ce112frow4_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_17c26732_d2b4_11ea_aff7_acbc32ce112frow4_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_17c26732_d2b4_11ea_aff7_acbc32ce112frow4_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_17c26732_d2b4_11ea_aff7_acbc32ce112frow4_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  cyan;\n",
       "        }</style><table id=\"T_17c26732_d2b4_11ea_aff7_acbc32ce112f\" ><thead>    <tr>        <th class=\"blank level0\" ></th>        <th class=\"col_heading level0 col0\" >inbound_text</th>        <th class=\"col_heading level0 col1\" >author_id</th>        <th class=\"col_heading level0 col2\" >created_at</th>        <th class=\"col_heading level0 col3\" >outbound_text</th>        <th class=\"col_heading level0 col4\" >response_tweet_id</th>    </tr></thead><tbody>\n",
       "                <tr>\n",
       "                        <th id=\"T_17c26732_d2b4_11ea_aff7_acbc32ce112flevel0_row0\" class=\"row_heading level0 row0\" >106643</th>\n",
       "                        <td id=\"T_17c26732_d2b4_11ea_aff7_acbc32ce112frow0_col0\" class=\"data row0 col0\" >anyone having issues with #osx #HighSierra and #Slack, #zoom or Multiple monitors? Getting major lag. @115858  @118190  @172377 #DevLife</td>\n",
       "                        <td id=\"T_17c26732_d2b4_11ea_aff7_acbc32ce112frow0_col1\" class=\"data row0 col1\" >AppleSupport</td>\n",
       "                        <td id=\"T_17c26732_d2b4_11ea_aff7_acbc32ce112frow0_col2\" class=\"data row0 col2\" >Wed Nov 22 01:02:29 +0000 2017</td>\n",
       "                        <td id=\"T_17c26732_d2b4_11ea_aff7_acbc32ce112frow0_col3\" class=\"data row0 col3\" >@823737 We're happy to help out with your concerns. To begin, have you made sure that all apps are up-to-date via the App Store &gt; Updates tab? \n",
       "Does the slower performance happen with a specific pattern?\n",
       "Send us a DM and we'll work on this together. https://t.co/GDrqU22YpT</td>\n",
       "                        <td id=\"T_17c26732_d2b4_11ea_aff7_acbc32ce112frow0_col4\" class=\"data row0 col4\" >2987500</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_17c26732_d2b4_11ea_aff7_acbc32ce112flevel0_row1\" class=\"row_heading level0 row1\" >106644</th>\n",
       "                        <td id=\"T_17c26732_d2b4_11ea_aff7_acbc32ce112frow1_col0\" class=\"data row1 col0\" >Hey @AppleSupport - not being able to duplicate a file in #pages when you searched for it is really annoying. What’s the fix?? https://t.co/CqnGyqNsLZ</td>\n",
       "                        <td id=\"T_17c26732_d2b4_11ea_aff7_acbc32ce112frow1_col1\" class=\"data row1 col1\" >AppleSupport</td>\n",
       "                        <td id=\"T_17c26732_d2b4_11ea_aff7_acbc32ce112frow1_col2\" class=\"data row1 col2\" >Wed Nov 22 02:33:58 +0000 2017</td>\n",
       "                        <td id=\"T_17c26732_d2b4_11ea_aff7_acbc32ce112frow1_col3\" class=\"data row1 col3\" >@689907 We're certainly glad to get you pointed in the right direction! Please reach out in DM and let us know which iPad model you're using and the version of iOS installed, as well as which country you're tweeting from. https://t.co/GDrqU22YpT</td>\n",
       "                        <td id=\"T_17c26732_d2b4_11ea_aff7_acbc32ce112frow1_col4\" class=\"data row1 col4\" >nan</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_17c26732_d2b4_11ea_aff7_acbc32ce112flevel0_row2\" class=\"row_heading level0 row2\" >106645</th>\n",
       "                        <td id=\"T_17c26732_d2b4_11ea_aff7_acbc32ce112frow2_col0\" class=\"data row2 col0\" >Yo @AppleSupport is that weird glitch w/ the capital “I️” your attempt to make my tweets less about me?  I️</td>\n",
       "                        <td id=\"T_17c26732_d2b4_11ea_aff7_acbc32ce112frow2_col1\" class=\"data row2 col1\" >AppleSupport</td>\n",
       "                        <td id=\"T_17c26732_d2b4_11ea_aff7_acbc32ce112frow2_col2\" class=\"data row2 col2\" >Wed Nov 22 02:29:26 +0000 2017</td>\n",
       "                        <td id=\"T_17c26732_d2b4_11ea_aff7_acbc32ce112frow2_col3\" class=\"data row2 col3\" >@823765 We'd love to help! Which device are you using? The latest versions of iOS include a fix for unexpected results with Auto-Correction while typing. If you haven't had a chance, we advise backing up and updating. Here's a guide to help: https://t.co/ahjigcvFRG</td>\n",
       "                        <td id=\"T_17c26732_d2b4_11ea_aff7_acbc32ce112frow2_col4\" class=\"data row2 col4\" >nan</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_17c26732_d2b4_11ea_aff7_acbc32ce112flevel0_row3\" class=\"row_heading level0 row3\" >106646</th>\n",
       "                        <td id=\"T_17c26732_d2b4_11ea_aff7_acbc32ce112frow3_col0\" class=\"data row3 col0\" >What the fuck @AppleSupport  my phone keeps hanging up calls and showing “call failure” 🤬😡</td>\n",
       "                        <td id=\"T_17c26732_d2b4_11ea_aff7_acbc32ce112frow3_col1\" class=\"data row3 col1\" >AppleSupport</td>\n",
       "                        <td id=\"T_17c26732_d2b4_11ea_aff7_acbc32ce112frow3_col2\" class=\"data row3 col2\" >Wed Nov 22 03:27:30 +0000 2017</td>\n",
       "                        <td id=\"T_17c26732_d2b4_11ea_aff7_acbc32ce112frow3_col3\" class=\"data row3 col3\" >@823779 We'd like to help. Send us a DM and we can start there. https://t.co/GDrqU22YpT</td>\n",
       "                        <td id=\"T_17c26732_d2b4_11ea_aff7_acbc32ce112frow3_col4\" class=\"data row3 col4\" >nan</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_17c26732_d2b4_11ea_aff7_acbc32ce112flevel0_row4\" class=\"row_heading level0 row4\" >106647</th>\n",
       "                        <td id=\"T_17c26732_d2b4_11ea_aff7_acbc32ce112frow4_col0\" class=\"data row4 col0\" >Is anyone having #iphone issues with there phone freezing randomly I have the 7 pulse updated to iOS 11.2 the newest version. Wtf is going on???? 🤬😡🤬😡🤬 @115858 @AppleSupport #whycan’tyouguysgetitright #stopmakingnewphonesuntillyoulearntomaketheoldonerunrite</td>\n",
       "                        <td id=\"T_17c26732_d2b4_11ea_aff7_acbc32ce112frow4_col1\" class=\"data row4 col1\" >AppleSupport</td>\n",
       "                        <td id=\"T_17c26732_d2b4_11ea_aff7_acbc32ce112frow4_col2\" class=\"data row4 col2\" >Wed Nov 22 04:15:29 +0000 2017</td>\n",
       "                        <td id=\"T_17c26732_d2b4_11ea_aff7_acbc32ce112frow4_col3\" class=\"data row4 col3\" >@823796 We'd love to offer our help in making sure your iPhone is working at peak performance. Send us a DM and let us know if you are using iOS 11.1.2 or iOS 11.2. You can confirm in Settings &gt; General &gt; About. https://t.co/GDrqU2kzhr</td>\n",
       "                        <td id=\"T_17c26732_d2b4_11ea_aff7_acbc32ce112frow4_col4\" class=\"data row4 col4\" >nan</td>\n",
       "            </tr>\n",
       "    </tbody></table>"
      ],
      "text/plain": [
       "<pandas.io.formats.style.Styler at 0x1928e04090>"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Showing the results of data wrangling\n",
    "show(apple_data)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Sweet. So now we are going to build off this `apple_data` from this point on.\n",
    "\n",
    "Since we are working with unsupervised learning for the first part of my chatbot, I do not need to do a train test split. My test set would also just end up needing labels as well. For this reason, I will just subset my data to the inbound/outbound data without employing the concept of unseen data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style  type=\"text/css\" >\n",
       "    #T_61deb2e4_d2b4_11ea_aff7_acbc32ce112frow0_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  cyan;\n",
       "        }    #T_61deb2e4_d2b4_11ea_aff7_acbc32ce112frow0_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_61deb2e4_d2b4_11ea_aff7_acbc32ce112frow0_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  cyan;\n",
       "        }    #T_61deb2e4_d2b4_11ea_aff7_acbc32ce112frow1_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  cyan;\n",
       "        }    #T_61deb2e4_d2b4_11ea_aff7_acbc32ce112frow1_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_61deb2e4_d2b4_11ea_aff7_acbc32ce112frow1_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  cyan;\n",
       "        }    #T_61deb2e4_d2b4_11ea_aff7_acbc32ce112frow2_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  cyan;\n",
       "        }    #T_61deb2e4_d2b4_11ea_aff7_acbc32ce112frow2_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_61deb2e4_d2b4_11ea_aff7_acbc32ce112frow2_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  cyan;\n",
       "        }    #T_61deb2e4_d2b4_11ea_aff7_acbc32ce112frow3_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  cyan;\n",
       "        }    #T_61deb2e4_d2b4_11ea_aff7_acbc32ce112frow3_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_61deb2e4_d2b4_11ea_aff7_acbc32ce112frow3_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  cyan;\n",
       "        }    #T_61deb2e4_d2b4_11ea_aff7_acbc32ce112frow4_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  cyan;\n",
       "        }    #T_61deb2e4_d2b4_11ea_aff7_acbc32ce112frow4_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_61deb2e4_d2b4_11ea_aff7_acbc32ce112frow4_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  cyan;\n",
       "        }</style><table id=\"T_61deb2e4_d2b4_11ea_aff7_acbc32ce112f\" ><thead>    <tr>        <th class=\"blank level0\" ></th>        <th class=\"col_heading level0 col0\" >Inbound</th>        <th class=\"col_heading level0 col1\" >Processed Inbound</th>        <th class=\"col_heading level0 col2\" >Outbound</th>    </tr></thead><tbody>\n",
       "                <tr>\n",
       "                        <th id=\"T_61deb2e4_d2b4_11ea_aff7_acbc32ce112flevel0_row0\" class=\"row_heading level0 row0\" >0</th>\n",
       "                        <td id=\"T_61deb2e4_d2b4_11ea_aff7_acbc32ce112frow0_col0\" class=\"data row0 col0\" >@AppleSupport The newest update. I️ made sure to download it yesterday.</td>\n",
       "                        <td id=\"T_61deb2e4_d2b4_11ea_aff7_acbc32ce112frow0_col1\" class=\"data row0 col1\" >['new', 'update', 'i️', 'make', 'sure', 'download', 'yesterday']</td>\n",
       "                        <td id=\"T_61deb2e4_d2b4_11ea_aff7_acbc32ce112frow0_col2\" class=\"data row0 col2\" >@115854 Lets take a closer look into this issue. Select the following link to join us in a DM and we'll go from there. https://t.co/GDrqU22YpT</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_61deb2e4_d2b4_11ea_aff7_acbc32ce112flevel0_row1\" class=\"row_heading level0 row1\" >1</th>\n",
       "                        <td id=\"T_61deb2e4_d2b4_11ea_aff7_acbc32ce112frow1_col0\" class=\"data row1 col0\" >@AppleSupport  https://t.co/NV0yucs0lB</td>\n",
       "                        <td id=\"T_61deb2e4_d2b4_11ea_aff7_acbc32ce112frow1_col1\" class=\"data row1 col1\" >['hey', 'anyone', 'else', 'upgraded', 'io', 'issue', 'capital', 'i️', 'mail', 'app', 'put']</td>\n",
       "                        <td id=\"T_61deb2e4_d2b4_11ea_aff7_acbc32ce112frow1_col2\" class=\"data row1 col2\" >@115854 We're here for you. Which version of the iOS are you running? Check from Settings &gt; General &gt; About.</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_61deb2e4_d2b4_11ea_aff7_acbc32ce112flevel0_row2\" class=\"row_heading level0 row2\" >2</th>\n",
       "                        <td id=\"T_61deb2e4_d2b4_11ea_aff7_acbc32ce112frow2_col0\" class=\"data row2 col0\" >@AppleSupport Tried resetting my settings .. restarting my phone .. all that</td>\n",
       "                        <td id=\"T_61deb2e4_d2b4_11ea_aff7_acbc32ce112frow2_col1\" class=\"data row2 col1\" >['hello', 'internet', 'someone', 'explain', 'symbol', 'keep', 'appear', 'phone', 'i️', 'try', 'type', 'letter', 'i️', 'also']</td>\n",
       "                        <td id=\"T_61deb2e4_d2b4_11ea_aff7_acbc32ce112frow2_col2\" class=\"data row2 col2\" >@115855 Let's go to DM for the next steps. DM us here: https://t.co/GDrqU22YpT</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_61deb2e4_d2b4_11ea_aff7_acbc32ce112flevel0_row3\" class=\"row_heading level0 row3\" >3</th>\n",
       "                        <td id=\"T_61deb2e4_d2b4_11ea_aff7_acbc32ce112frow3_col0\" class=\"data row3 col0\" >@AppleSupport This is what it looks like https://t.co/XCQU2l4xUB</td>\n",
       "                        <td id=\"T_61deb2e4_d2b4_11ea_aff7_acbc32ce112frow3_col1\" class=\"data row3 col1\" >['get', 'screenshot', 'say', 'iphonex', 'reserve', 'email', 'say', 'happen']</td>\n",
       "                        <td id=\"T_61deb2e4_d2b4_11ea_aff7_acbc32ce112frow3_col2\" class=\"data row3 col2\" >@115855 Any steps tried since it started last night?</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_61deb2e4_d2b4_11ea_aff7_acbc32ce112flevel0_row4\" class=\"row_heading level0 row4\" >4</th>\n",
       "                        <td id=\"T_61deb2e4_d2b4_11ea_aff7_acbc32ce112frow4_col0\" class=\"data row4 col0\" >@AppleSupport I️ have an iPhone 7 Plus and yes I️ do</td>\n",
       "                        <td id=\"T_61deb2e4_d2b4_11ea_aff7_acbc32ce112frow4_col1\" class=\"data row4 col1\" >['thank', 'update', 'phone', 'even', 'slow', 'barely', 'work', 'thank', 'ruin', 'phone']</td>\n",
       "                        <td id=\"T_61deb2e4_d2b4_11ea_aff7_acbc32ce112frow4_col2\" class=\"data row4 col2\" >@115855 That's great it has iOS 11.1 as we can rule out being outdated. Any steps tried since this started? Do you recall when it started?</td>\n",
       "            </tr>\n",
       "    </tbody></table>"
      ],
      "text/plain": [
       "<pandas.io.formats.style.Styler at 0x1928e046d0>"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Just making for the nice display\n",
    "show = lambda x: x.head(5).style.set_properties(**{'background-color': 'black',                                                   \n",
    "                                    'color': 'lawngreen',                       \n",
    "                                    'border-color': 'white'})\\\n",
    ".applymap(lambda x: f\"color: {'cyan' if isinstance(x,str) else 'lawngreen'}\")\\\n",
    ".background_gradient(cmap='Blues')\n",
    "\n",
    "# Reading in processed inbound\n",
    "processed_inbound = pd.read_pickle('objects/processed_inbound.pkl')\n",
    "\n",
    "\n",
    "# Displaying processed_text\n",
    "display = pd.DataFrame({\"Inbound\": inbound, \"Processed Inbound\": \n",
    "                        list(processed_inbound) + [''] * (len(inbound) - len(processed_inbound)),\n",
    "                       \"Outbound\": outbound})\n",
    "\n",
    "show(display)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This outbound data will be useful for natural langauge generation, and I might want to store the id where it points to as well.\n",
    "\n",
    "When we process this outbound data for natural langauge generation, would this have a different preprocessing steps? Would we want it to have more complete sentences so maybe we would not try to remove the actual stop words and punctuation."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Text Preprocessing Pipeline\n",
    "In this process, it's expected that I will find much edge cases, and I don't have to address them all. However, in the end, I felt like my preprocessing pipeline is relatively complete given this dataset.\n",
    "\n",
    "Also, this was an iterative process because I fit my models in the next notebook just to notice that my Tweets still had '\\u200d', '#', '', and '@' even though I still did a lot of fitting already. These are just emojis and extra punctuation I decided at a later stage to remove, but in the end I just modified my pipeline function. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Before that, I did a bit of EDA with bag of words to see what words are being used the most by the dataset to better preprocess the data and prevent working in too much granularity. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Writing a function to plot the counts that are in the data\n",
    "\n",
    "def top10_bagofwords(data, output_name, title):\n",
    "    ''' Taking as input the data and plots the top 10 words based on counts in this text data'''\n",
    "    bagofwords = CountVectorizer()\n",
    "    inbound = bagofwords.fit_transform(data)\n",
    "    inbound # Output will be a sparse matrix\n",
    "    # Inspecting of often contractions and colloquial language is used\n",
    "    word_counts = np.array(np.sum(inbound, axis=0)).reshape((-1,))\n",
    "    words = np.array(bagofwords.get_feature_names())\n",
    "    words_df = pd.DataFrame({\"word\":words, \n",
    "                             \"count\":word_counts})\n",
    "    words_rank = words_df.sort_values(by=\"count\", ascending=False)\n",
    "    # words_rank.to_csv('words_rank.csv') # Storing it in a csv so I can inspect and go through it myself\n",
    "    words_rank.head()\n",
    "    # Visualizing top 10 words\n",
    "    plt.figure(figsize=(12,6))\n",
    "    sns.barplot(words_rank['word'][:10], words_rank['count'][:10].astype(str), palette = 'inferno')\n",
    "    plt.title(title)\n",
    "    \n",
    "    # Saving\n",
    "    plt.savefig(f'visualizations/{output_name}.png')\n",
    "    \n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAuIAAAGECAYAAABzpkXlAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOzdfVyUVf7/8ffIjGhiES4Isa5rW+qupPSVMjIhrRUISCPdSsraWtvuzGyjUEjCzdWKr5oautuWu1lbkRq4LmI3pm5hrVqb8c1utlVDNG4UESlgGM7vD38MIqBoM16Jr+fj0QPnzHXzOYcZes+ZM9fYjDFGAAAAAE6pLlYXAAAAAJyJCOIAAACABQjiAAAAgAUI4gAAAIAFCOIAAACABQjiAAAAgAUI4sBJePzxxzVmzBiNGTNGYWFhiomJcd+ura31yDk+++wzXXHFFS3aPv74Y11//fWKi4vTrbfeqrKysjb3HTBggEaNGqWjr066cOFCDRgwQJ988slJ1VRcXKzJkye3e/9XX32lyZMnKzExUddee61uvvlmbdmy5aTOZaV//OMfGjt2bIu2G264QSNGjGgxppMmTdLf/va3kz7PJ598olGjRnV4+9LSUt14440ndI6VK1dqwIABWrBgQYt2Y4yuuuoqJSQknNDxbrnlFhUUFBxzmw8++OCEj+spv/3tb7Vy5cpW7QsXLtRll13mfp7Gx8frwQcf1M6dO497zOrqak2cOPGEaxk1apT7b0NiYqLi4+P16quvdmjfjIwMjRo1SvPmzTvh83bEgAEDlJiYqDFjxmjs2LFKSEjQH//4R6+c64fg5Zdf1p/+9CerywBasVtdAHA6Sk9Pd/971KhRysrK0kUXXeSRYzc0NOjFF1/Us88+q2+//dbdXl9fr/vvv19z587V0KFD9be//U1paWl69tln2zyOMUZbtmzRJZdc4r69Zs0anXPOOSdd2549e7Rjx4427/vvf/+rW2+9VbNnz9aIESMkSZs2bdJdd92ll19+WRdeeOFJn/dUGz58uB555BEdOHBA/v7+2r9/v8rKytSrVy998sknGjx4sBoaGrR161Y99thjp6yu3r1765VXXjnh/c477zytWrVK999/v7tty5Ytqq2tVffu3T1Z4g/aNddcoxkzZrhv5+bm6tZbb9U//vEP+fn5tbtfVVXVSb94PfJvw969exUTE6OoqCiFhIQcc79XX31V69evV3Bw8EmdtyP++te/KiAgQJJ06NAhjRkzRv3799fIkSO9dk6r3HTTTVaXALSJIA542DPPPKN//OMf8vHxUb9+/fToo48qMDBQt9xyi37xi19o69atqqys1JgxY1oEoyaffvqpPv/8cy1atEi33367u/2TTz6Rn5+fhg4dKkkaN26c/vCHP6iyslLnnntuq+Nce+21WrVqlTuIb926VRdccEGLGfu33npLixYtUmNjo3r06KFp06Zp8ODB+uqrr5SWlqb6+noZYzRu3DjdeOONSk9PV2lpqe644w4999xzLc737LPP6vrrr3eHcEmKjIzU//7v/6pbt27HPN/ChQv19ddfq7S0VOXl5Ro0aJCGDRum3Nxc7d69WykpKUpISOjwdk6nU3PmzNGmTZvk4+OjwYMHa9q0afLz89OoUaN03XXXadOmTdq7d6/GjBmjBx54oEVf/P39FRYWpi1btujqq6/W+vXrdcUVV6hXr15at26dBg8erI8//lihoaEKDQ3VN998o8cee0wlJSUyxmjs2LH6zW9+o927dys5OVk/+9nPVFJSomXLlumtt97SX//6V/n5+al///7uc7Y15snJyS3q2r17txITE/XRRx9p4cKFKikpUXl5uUpKStS7d2899dRTCgoKavVY6N+/v/bu3asPP/xQ//M//yNJev3113Xttdfqn//8p4wxio2N1YwZMzR8+HBJUlpamvr3769bb7211fGaHGssv/32W91///3atWuXzj77bM2cOVP9+vVTdXW1MjMz9dlnn8lms2nEiBF68MEHZbfbNWDAAG3atMkdDptuf/nll5o3b5769OmjL7/8Ug0NDcrMzNTQoUNVWlqq1NRUlZWV6bzzztO+ffvarfdoY8eO1apVq/T3v/9dN910k5YvX65XX31VTqdTVVVVmjRpkiZMmKBp06aptrZWY8aM0cqVK/X666+3ud3xVFVVqXv37jrrrLMkHf6dz5o1SwcOHJDL5dItt9yicePGacKECTLGaNKkScrIyNA555yjmTNn6sCBA7LZbLr99ts1duxYffDBB5o1a5bOOuss1dTUaMWKFXr33Xe1ePFiOZ1OdevWTY888oguvvji49bm5+ensLAw/fe//9VZZ53V4eMuXLhQu3bt0jfffKPy8nINHDhQs2bNcj/XBg8erM8//1wPPvigfvrTn7bZD0lavny5li5dqi5duujcc8/VE088oZCQEK1bt67N87b3fGmvfeHChaqsrNSMGTOO+bj905/+pOXLl6tHjx6KiIjQ22+/rXXr1nX4MQWcMAPgexk5cqTZtm2bMcaY5cuXmxtuuMHU1NQYY4xZsGCBuf32240xxtx8881m0qRJpr6+3lRVVZmYmBizbt26do9bXFxswsPD3bdXr17tPlaTESNGmO3bt7fat3///uaLL74ww4YNM3V1dcYYY6ZPn27WrVvnrvc///mPufzyy83XX39tjDGmsLDQDB8+3FRXV5tp06aZP/7xj8YYY8rKyswDDzxgXC6Xef/99018fHyb9SYkJJj169e3259jnW/BggVm5MiR5uDBg+a7774zl1xyiZk9e7Yxxpg333zTjB492j2eHdnu6aefNvfdd5+pr683LpfLpKammkcffdQYc/j3NWfOHGOMMd9884256KKL3DUdadGiRebxxx83xhhz3333mbfeest89NFHJjEx0RhjzMKFC81TTz1ljDEmOTnZPP/888YYYw4ePGgSExPN6tWrTXFxsenfv7/ZvHmzMcaYTz/91ERGRpqysjJjjDGPPvqoGTlypDHGtDvmRzryMbFgwQJz1VVXmerqamOMMb/97W/N008/3aofK1asMHfeead57rnnzIwZM4wxxnz77bdm9OjR5r333nP/PpcuXWruv/9+Y4wx1dXV5rLLLjNVVVWtjnfzzTebNWvWHHMs33//fTNw4ECzdetWY4wxr7zyihk3bpwxxpiHH37Y/P73vzeNjY2mrq7O3H777e5+9+/f3+zbt899rqbb77//vvn5z39uPv30U2OMMc8995xJTk42xhhzzz33mHnz5hljjNm5c6cJDw83K1asaFX3ggULTGZmZqv2OXPmmMcee8wcOnTI/OpXvzL79+83xhjz0Ucfucf6yHE/1nZHGzlypBk9erS59tprTUxMjBk4cKCZO3euMcYYp9NprrnmGlNUVGSMOfy4iYuLMx999FGLvjudTnPVVVeZtWvXusd5xIgR5sMPP3SP8+7du40xxuzYscMkJCS4a/viiy/M8OHD3X+PjnT0WH/11VcmMjLSfPzxxyd03AULFpioqChTXl5uXC6XefDBB92PiZEjR5pFixa5+9teP7Zv326GDRtm9uzZY4w5/Fh89NFHj3ne9p4v7bUf+ftv73G7ceNGExMTY6qqqkxjY6OZNm2a+/kJeAsz4oAHbdy4UUlJSe4Zr4kTJ2rJkiWqr6+XdHidscPhkMPhUGxsrN59990Ovw3c2Ngom83Wos0YIx8fnza379WrlwYPHqx33nlH0dHR2rJlizIzM933v//++7rsssvUp08fSYdnrwMCAlRUVKRf/vKXeuSRR7Rt2zZFRkYqPT1dXboc+yMlNptNjY2N7d5/rPNJ0uWXX66ePXtKkoKCgtwz6z/5yU904MAB93E6st3GjRs1depUORwOSYfXNd97773uY1x11VWSDi/16NWrl6qqqtx1NYmKilJ6errq6+u1ZcsWPfnkk/L19VVFRYUqKir0wQcfaMqUKfr222/14Ycf6vnnn5ck9ezZU0lJSdq4caOGDBkiu92u8PBwSYeX6gwfPlyBgYGSDj8e3n33XUk6qTG/9NJL3UsqfvGLX6iqqqrdbZvWA6elpenNN9/UqFGjWjx2kpKS9Mwzz2j//v0qKCjQlVdeqbPPPvuY529vLKXDs9lNs+/XXXedHnvsMVVXV2vjxo16+eWXZbPZ1LVrV914443661//qjvvvPOY5znvvPP085//3N3X119/XZJUWFioRx55RJLUt29fDRs27Lg1H8lms6lbt27q0aOHlixZog0bNmjnzp367LPPWiwNa9LR7ZocuTSluLhYt912my688EINHDhQX3/9taZPn+7etra2Vp9++qn78SJJO3fuVF1dnUaPHi3p8DiPHj1a//znPzVs2DCFhIQoNDRUkvTee++prKxMt912W4v+ff311xo4cGCr2m699VZ16dJFjY2N6t69ux5++GENHjxYH3zwQYePK0mxsbH60Y9+JKn5nbqm30lERMRx+9GzZ09dccUV7uU6Ted56aWX2j1ve8+Xjj6P2nrcbtiwQbGxse7HfXJyst5///22f7GAhxDEAQ86Oiw3NjaqoaHBfdtub37KGWOOG7SOFBIS0uLDmU6nUwcOHFDv3r3b3afprff6+nqNGjWqxfnbC/YNDQ0aOXKk1q5dq8LCQm3atEnPPPNMmx+AO1J4eLj+/e9/t3phsWjRIv3kJz855vkkqWvXri3uO7LWI3Vku7Z+D06n033b19fX/W+bzdbqQ62SFBYWpn379umtt95SWFiYey31iBEj9N5772nHjh0KDw9XbW1tq/2P/L137dq11e+9yZFBuL0xP9Ya4aYlP8fqR5PAwED94he/0MaNG5Wbm6vU1FRVVla67z/77LMVGxvrXqqRkZHR7rGO1N5YHv3Yttlsstvtx32ONGl68Xq8vh7d7/YeN+355JNPdP311+ubb77RDTfcoF/96lcaOnSoYmNj9c4777TavqPbtaVPnz4aNWqUNm/erAsvvFA9e/ZUXl6e+/6Kigr3i8wmLpfrmM+bphf90uGxjIyM1Pz5891te/fubXO5ktRyjfjROnrcN998s8XjuLGxscXvvuk4x+qHj49Pi/tqa2tVUlJyzPMOHDiwzedLR/92tfW4tdvt7T4/AW/hqimAB40YMUIrVqxwz5AtW7ZMl1xyiTs8rlq1So2NjaqqqtKaNWtO6IoZQ4YM0YEDB/Thhx9KklasWKHw8PBjzlpeddVV+uijj/TSSy/puuuua3FfZGSk3n33XRUXF0uSe73kkCFD9Lvf/U75+fmKj49XRkaG/Pz89PXXX8vHx6dFoD3SHXfcoddee809wysdnpletmyZBg4ceMzzedqIESP08ssvy+l0qrGxUS+99JJ77XNH2Ww2XX755VqyZImuvPJKd/uVV16p559/Xpdeeqnsdrv8/Pw0ZMgQvfTSS5IOX2EjNzdXl19+eatjDh8+XO+9956++eYbSXLP6kpqd8w9aezYsVq6dKmqq6tbrE9vkpycrBdeeEHGGA0ePPh7nevzzz/X9u3bJR3+4OHQoUPVvXt3XXHFFXrxxRdljFF9fb1ycnLcYxUQEOD+UOTq1as7dJ4RI0a4r0SyZ88effDBBx2u8bXXXtPu3bsVFxenoqIiBQQE6J577tEVV1zhDtcul0t2u10ul0vGmGNudzzffvutNm/erMGDB6tfv37q1q2bO4jv3btXCQkJ7neImpx//vmy2+164403JB2+cs7atWvbfHxFRkbqvffe01dffSVJ2rBhg6699trvfSWn4x337bffVnV1tRobG5WTk9Pmu3zH6sewYcO0adMm90TDK6+8oqeeeuqY523v+fJ9nkfR0dF64403VF1dLenwunXA25gRBzxo3Lhx2rt3r8aPH6/Gxkb17dtXWVlZ7vtra2s1btw41dTUaMKECYqMjOzwsR0OhxYtWqSZM2fqu+++k7+/v5544olj7uPr66tRo0bp008/bRW8LrjgAmVkZOi+++6Ty+VSt27dtGTJEvXs2VP33HOP0tLS9Oqrr8rHx0dXX321LrnkElVVVcnX11fjxo3Ta6+91mIWq2/fvlqyZInmz5+vJ554Qo2NjQoICNDixYvd527vfJ52991364knntDYsWPV0NCgwYMH69FHHz3h40RFRSkvL69FsLjiiiuUkpKiX//61+62rKwszZw5UytXrlR9fb0SExOVlJSkkpKSFscbMGCAUlJSdOutt6pHjx4twm57Y+5JV199tTIyMjR16tQ27x84cKDOOeecE75EYlvOP/98LVq0SMXFxerVq5fmzJkj6fAVhx5//HElJibK6XRqxIgRuuuuu9z3zZw5U2effbYuv/xy9xKeY8nIyNC0adMUFxen4ODgNpdgNMnPz9fWrVvdy6j69eunF154Qb6+vho+fLiWL1+u2NhY2Ww2XXrppQoICNCuXbvUt29fDR48WPHx8Vq6dKl69+7d5nbnn39+q3M+9NBD6tatm2w2m7777jvFxcXp+uuvlyRlZ2dr1qxZ+vOf/6yGhgZNmTLF/WHsJg6HQ9nZ2Xr88ce1cOFCuVwu3Xvvvbrssstavei44IILNHPmTD344IPuGd7FixerR48exx3HYznecX/0ox9p0qRJqqys1CWXXOL+fXa0H5KUkpKi3/zmN5IOv3vzhz/8Qb179273vO09X3r16tVme0deoEVGRupXv/qVbrjhBnXr1k0XXnjhGXVVIVjDZo71XiYAj7nllluUnJys2NhYq0sB2vT111+7rxNOAEFHHHk1ktPdJ598oo8++sh9zfilS5fq448/brE0BvA0ZsQBAHr66aeVk5OjzMxMQjjOSP369dOzzz6rnJwc2Ww2hYSE6Pe//73VZaGTY0YcAAAAsAAf1gQAAAAsQBAHAAAALHDGrRFvbGxUTU2NHA5Hq2uaAgAAAJ5ijJHT6VSPHj3a/O6QMy6I19TU6IsvvrC6DAAAAJwh+vfv3+bles+4IN70ldf9+/dv9Q19AAAAgKfU19friy++cOfPo51xQbxpOUrXrl1bfMUtAAAA4A3tLYfmw5oAAACABQjiAAAAgAUI4gAAAIAFCOIAAACABQjiAAAAgAW8GsTz8vIUHx+v+Ph4PfHEE5KkwsJCJSYmavTo0Zo3b5572+3btyspKUkxMTFKS0tTQ0ODJGnPnj1KTk5WbGys7r77btXU1EiSDh48qDvvvFNxcXFKTk5WeXm5N7sCAAAAeJTXgvh3332nWbNmadmyZcrLy9OWLVu0bt06TZ8+XdnZ2crPz1dRUZE2bNggSUpJSdGMGTO0du1aGWOUk5MjScrMzNSECRNUUFCgsLAwZWdnS5Lmz5+viIgIrVmzRuPHj9esWbO81RUAAADA47wWxF0ulxobG/Xdd9+poaFBDQ0N8vPzU9++fdWnTx/Z7XYlJiaqoKBAJSUlqq2tVXh4uCQpKSlJBQUFcjqd2rx5s2JiYlq0S9L69euVmJgoSUpISNDGjRvldDq91R0AAADAo7z2hT5+fn6aMmWK4uLi1L17d11yySUqKytTYGCge5ugoCCVlpa2ag8MDFRpaakqKyvl5+cnu93eol1Si33sdrv8/Py0f/9+9e7d21tdAgAAADzGa0H8s88+04oVK/TOO++oZ8+eeuihh7Rz584W3yxkjJHNZlNjY2Ob7U0/j9TeNxMZY9SlS8cn+IuKik6wRwAAAIDneC2Iv/vuu4qMjFSvXr0kHV5W8txzz8nHx8e9TXl5uYKCghQcHNziw5YVFRUKCgpSQECAqqur5XK55OPj495eOjybXlFRoeDgYDU0NKimpkb+/v4dri8sLIyvuAcAAIDX1NXVHXPy12trxAcOHKjCwkJ9++23MsZo3bp1GjJkiHbs2KFdu3bJ5XJp9erVioqKUmhoqHx9fbV161ZJh6+2EhUVJYfDoYiICOXn50uScnNzFRUVJUmKjo5Wbm6uJCk/P18RERFyOBze6g4AAADgUTZjjPHWwf/0pz9p5cqVcjgcuuiii5SRkaEPP/xQs2fPVl1dnaKjozVt2jTZbDZ99tlnSk9P16FDhzRo0CDNnj1bXbt2VUlJiVJTU7Vv3z6FhIRo7ty5Ouecc3TgwAGlpqaquLhYPXv2VFZWln784x8ft6amVybMiAMAAMCbjpc7vRrEf4iONyC1tXXq1q3zBfTO2i8AAIAfquPlTq+tET9ddevmq+Bzw60uw+O+qfy31SUAAADgCHzFPQAAAGABgjgAAABgAYI4AAAAYAGCOAAAAGABgjgAAABgAYI4AAAAYAGCOAAAAGABgjgAAABgAYI4AAAAYAGCOAAAAGABgjgAAABgAYI4AAAAYAGCOAAAAGABgjgAAABgAYI4AAAAYAGCOAAAAGABgjgAAABgAYI4AAAAYAGCOAAAAGABgjgAAABgAYI4AAAAYAGCOAAAAGABgjgAAABgAYI4AAAAYAGCOAAAAGABgjgAAABgAYI4AAAAYAGCOAAAAGABgjgAAABgAbu3Dvzaa6/pxRdfdN/evXu3xowZo6uvvlqzZ89WXV2d4uLiNHXqVEnS9u3blZaWppqaGkVERCgzM1N2u1179uxRSkqK9u3bp379+ikrK0s9evTQwYMH9dBDD6m4uFgBAQGaP3++AgMDvdUdAAAAwKO8NiM+fvx45eXlKS8vT1lZWerVq5cmTZqk6dOnKzs7W/n5+SoqKtKGDRskSSkpKZoxY4bWrl0rY4xycnIkSZmZmZowYYIKCgoUFham7OxsSdL8+fMVERGhNWvWaPz48Zo1a5a3ugIAAAB43ClZmvLYY49p6tSpKi4uVt++fdWnTx/Z7XYlJiaqoKBAJSUlqq2tVXh4uCQpKSlJBQUFcjqd2rx5s2JiYlq0S9L69euVmJgoSUpISNDGjRvldDpPRXcAAACA783rQbywsFC1tbWKi4tTWVlZi+UjQUFBKi0tbdUeGBio0tJSVVZWys/PT3a7vUW7pBb72O12+fn5af/+/d7uDgAAAOARXlsj3uSVV17Rr3/9a0lSY2OjbDab+z5jjGw2W7vtTT+PdPTtI/fp0qXjryuKiorabB86dGiHj3G62bp1q9UlAAAA4P/zahCvr6/X5s2bNWfOHElScHCwysvL3feXl5crKCioVXtFRYWCgoIUEBCg6upquVwu+fj4uLeXDs+mV1RUKDg4WA0NDaqpqZG/v3+HawsLC5Ovr6+Henp66MwvMgAAAH5o6urq2p38lby8NOXzzz/XT3/6U5111lmSpCFDhmjHjh3atWuXXC6XVq9eraioKIWGhsrX19c9Y5uXl6eoqCg5HA5FREQoPz9fkpSbm6uoqChJUnR0tHJzcyVJ+fn5ioiIkMPh8GZ3AAAAAI/x6ox4cXGxgoOD3bd9fX01Z84cTZ48WXV1dYqOjlZsbKwkKSsrS+np6Tp06JAGDRqkiRMnSpIyMjKUmpqqxYsXKyQkRHPnzpUkTZkyRampqYqPj1fPnj2VlZXlza4AAAAAHmUzxhiriziVmt4iONbSlOBzw09xVd73TeW/rS4BAADgjHK83Mk3awIAAAAWIIgDAAAAFiCIAwAAABYgiAMAAAAWIIgDAAAAFiCIAwAAABYgiAMAAAAWIIgDAAAAFiCIAwAAABYgiAMAAAAWIIgDAAAAFiCIAwAAABYgiAMAAAAWIIgDAAAAFiCIAwAAABYgiAMAAAAWIIgDAAAAFiCIAwAAABYgiAMAAAAWIIgDAAAAFiCIAwAAABYgiAMAAAAWIIgDAAAAFiCIAwAAABYgiAMAAAAWIIgDAAAAFiCIAwAAABYgiAMAAAAWIIgDAAAAFiCIAwAAABYgiAMAAAAW8GoQX7dunZKSkhQXF6fHH39cklRYWKjExESNHj1a8+bNc2+7fft2JSUlKSYmRmlpaWpoaJAk7dmzR8nJyYqNjdXdd9+tmpoaSdLBgwd15513Ki4uTsnJySovL/dmVwAAAACP8loQLy4uVkZGhrKzs7Vq1Sp9+umn2rBhg6ZPn67s7Gzl5+erqKhIGzZskCSlpKRoxowZWrt2rYwxysnJkSRlZmZqwoQJKigoUFhYmLKzsyVJ8+fPV0REhNasWaPx48dr1qxZ3uoKAAAA4HFeC+JvvvmmrrnmGgUHB8vhcGjevHnq3r27+vbtqz59+shutysxMVEFBQUqKSlRbW2twsPDJUlJSUkqKCiQ0+nU5s2bFRMT06JdktavX6/ExERJUkJCgjZu3Cin0+mt7gAAAAAeZffWgXft2iWHw6G77rpLe/fu1ZVXXqkLL7xQgYGB7m2CgoJUWlqqsrKyFu2BgYEqLS1VZWWl/Pz8ZLfbW7RLarGP3W6Xn5+f9u/fr969e3eovqKiojbbhw4delL9PR1s3brV6hIAAADw/3ktiLtcLm3ZskXLli3TWWedpbvvvlvdunWTzWZzb2OMkc1mU2NjY5vtTT+PdPTtI/fp0qXjE/xhYWHy9fU9wV6d3jrziwwAAIAfmrq6unYnfyUvLk350Y9+pMjISAUEBKhbt266+uqrVVhY2OJDleXl5QoKClJwcHCL9oqKCgUFBSkgIEDV1dVyuVwttpcOz6ZXVFRIkhoaGlRTUyN/f39vdQcAAADwKK8F8ZEjR+rdd9/VwYMH5XK59M9//lOxsbHasWOHdu3aJZfLpdWrVysqKkqhoaHy9fV1L53Iy8tTVFSUHA6HIiIilJ+fL0nKzc1VVFSUJCk6Olq5ubmSpPz8fEVERMjhcHirOwAAAIBH2YwxxlsHX758uf7yl7/I6XRq+PDhSk9P1wcffKDZs2errq5O0dHRmjZtmmw2mz777DOlp6fr0KFDGjRokGbPnq2uXbuqpKREqamp2rdvn0JCQjR37lydc845OnDggFJTU1VcXKyePXsqKytLP/7xj49bU9NbBMdamhJ8brinh8Jy31T+2+oSAAAAzijHy51eDeI/RARxAAAAnArHy518syYAAABgAYI4AAAAYAGCOAAAAGABgjgAAABgAYI4AAAAYAGCOAAAAGABgjgAAABgAYI4AAAAYAGCONpVV1tvdQle0Vn7BQAATi92qwvAD5dvt666PPAmq8vwuMLyl60uAcEj3p4AACAASURBVAAAgBlxAAAAwAoEcQAAAMACBHEAAADAAgRxAAAAwAIEcaAD6mudVpfgFZ21XwAAnA64agrQAV27OXRj399YXYbHvbLrz1aXAADAGYsZcQAAAMACBHEAAADAAgRxAAAAwAIEcQAAAMACBHEAAADAAgRxAAAAwAIEcQAAAMACBHEAAADAAgRxAAAAwAIEcQAAAMACBHEAAADAAgRxAAAAwAIEcQAAAMACBHEAAADAAnZvHvyWW27R/v37ZbcfPs3MmTNVU1Oj2bNnq66uTnFxcZo6daokafv27UpLS1NNTY0iIiKUmZkpu92uPXv2KCUlRfv27VO/fv2UlZWlHj166ODBg3rooYdUXFysgIAAzZ8/X4GBgd7sDgAAAOAxXpsRN8Zo586dysvLc/83YMAATZ8+XdnZ2crPz1dRUZE2bNggSUpJSdGMGTO0du1aGWOUk5MjScrMzNSECRNUUFCgsLAwZWdnS5Lmz5+viIgIrVmzRuPHj9esWbO81RUAAADA47wWxP/73/9Kkm6//XZde+21evHFF7Vt2zb17dtXffr0kd1uV2JiogoKClRSUqLa2lqFh4dLkpKSklRQUCCn06nNmzcrJiamRbskrV+/XomJiZKkhIQEbdy4UU6n01vdAQAAADzKa0H84MGDioyM1DPPPKO//OUveuWVV7Rnz54Wy0eCgoJUWlqqsrKyFu2BgYEqLS1VZWWl/Pz83EtbmtoltdjHbrfLz89P+/fv91Z3AAAAAI/y2hrxiy++WBdffLH79rhx47RgwQINHTrU3WaMkc1mU2Njo2w2W6v2pp9HOvr2kft06dLx1xVFRUVtth9ZX2ezdevWE9qesWjGWAAAAE/zWhDfsmWLnE6nIiMjJR0OyqGhoSovL3dvU15erqCgIAUHB7dor6ioUFBQkAICAlRdXS2XyyUfHx/39tLh2fSKigoFBweroaFBNTU18vf373B9YWFh8vX19VBvTw+dOUyeKMaiGWMBAIB31NXVtTv5K3lxaUp1dbWefPJJ1dXV6dChQ3r99df14IMPaseOHdq1a5dcLpdWr16tqKgohYaGytfX1z0zl5eXp6ioKDkcDkVERCg/P1+SlJubq6ioKElSdHS0cnNzJUn5+fmKiIiQw+HwVncAAAAAj/LajPjIkSP18ccfa+zYsWpsbNSECRN08cUXa86cOZo8ebLq6uoUHR2t2NhYSVJWVpbS09N16NAhDRo0SBMnTpQkZWRkKDU1VYsXL1ZISIjmzp0rSZoyZYpSU1MVHx+vnj17Kisry1tdAQAAADzOZowxVhdxKjW9RXCspSnB54af4qq875vKf5/UfpcH3uThSqxXWP7ySe13Y9/feLgS672y689WlwAAQKd1vNzJN2sCOCHOus53mdDO2CcAwA+fV79ZE0Dn4/B1aPrgO60uw6P+sO1PVpcAADgDMSMOAAAAWIAgDgAAAFiAIA4AAABYgCAOAAAAWIAgDgAAAFiAIA4AAABYgCAOAAAAWIAgDgAAAFiAIA4AAABYgCAOAAAAWIAgDgAAAFiAIA4AAABYgCAOAAAAWKBDQby0tLRV23/+8x+PFwMAAACcKY4ZxA8cOKADBw5o0qRJqqqqct+uqKjQfffdd6pqBAAAADod+7Hu/N3vfqf33ntPkjRs2LDmnex2xcTEeLcyAAAAoBM7ZhB/7rnnJEnTpk3T7NmzT0lBAAAAwJngmEG8yezZs1VSUqKqqioZY9ztgwYN8lphAAAAQGfWoSC+YMECPffcc+rVq5e7zWaz6e233/ZaYQAAAEBn1qEgnpubqzfeeEO9e/f2dj0AAADAGaFDly8MCQkhhAMAAAAe1KEZ8cjISD355JO66qqr1K1bN3c7a8QBAACAk9OhIL5y5UpJUkFBgbuNNeIAAADAyetQEF+3bp236wAAAADOKB0K4kuXLm2z/de//rVHiwEAAADOFB0K4l988YX73/X19dq8ebMiIyO9VhQAAADQ2XX4C32OVFpaqrS0NK8UBACni4Y6p+y+DqvL8KjO2CcA+KHqUBA/Wu/evVVSUuLpWgDgtGL3dejZ6DusLsOjJm14zuoSAOCMccJrxI0xKioqavEtm8fyxBNPqLKyUnPmzFFhYaFmz56turo6xcXFaerUqZKk7du3Ky0tTTU1NYqIiFBmZqbsdrv27NmjlJQU7du3T/369VNWVpZ69OihgwcP6qGHHlJxcbECAgI0f/58BQYGnkT3AQAAAGt06At9vvjiC/d/X375pUJCQpSVlXXc/TZt2qTXX39dklRbW6vp06crOztb+fn5Kioq0oYNGyRJKSkpmjFjhtauXStjjHJyciRJmZmZmjBhggoKChQWFqbs7GxJ0vz58xUREaE1a9Zo/PjxmjVr1kl1HgAAALBKh4L47NmzNXv2bN1333266667lJKSouDg4GPuc+DAAc2bN0933XWXJGnbtm3q27ev+vTpI7vdrsTERBUUFKikpES1tbUKDw+XJCUlJamgoEBOp1ObN29WTExMi3ZJWr9+vRITEyVJCQkJ2rhxo5xO58mNAAAAAGCBDi1N2bVrl+655x6VlZWpsbFR5557rv74xz/qZz/7Wbv7zJgxQ1OnTtXevXslSWVlZS2WjwQFBam0tLRVe2BgoEpLS1VZWSk/Pz/Z7fYW7Ucfy263y8/PT/v371fv3r1PsPsAAACANToUxGfOnKnf/OY3uu666yRJK1asUGZmpl544YU2t3/ttdcUEhKiyMhI97dyNjY2ymazubcxxshms7Xb3vTzSEffPnKfLl06NLnvVlRU1Gb70KFDT+g4p5OtW7ee0PaMRTPGollnHYsTHQeJsQAAfD8dCuL79u1zh3BJuv766/WXv/yl3e3z8/NVXl6uMWPGqKqqSt9++61KSkrk4+Pj3qa8vFxBQUEKDg5WeXm5u72iokJBQUEKCAhQdXW1XC6XfHx83NtLh2fTKyoqFBwcrIaGBtXU1Mjf3/+EOh4WFiZfX98T2ud011lDw8lgLJoxFocxDs0YCwDwjLq6unYnf6UOrhF3uVw6cOCA+/b+/fuPuf3SpUu1evVq5eXl6f7779eoUaP05z//WTt27NCuXbvkcrm0evVqRUVFKTQ0VL6+vu4ZmLy8PEVFRcnhcCgiIkL5+fmSpNzcXEVFRUmSoqOjlZubK+lw6I+IiJDDwXVvAQAAcPro0Iz4zTffrBtuuEFxcXGy2WzKz8/XrbfeekIn8vX11Zw5czR58mTV1dUpOjpasbGxkqSsrCylp6fr0KFDGjRokCZOnChJysjIUGpqqhYvXqyQkBDNnTtXkjRlyhSlpqYqPj5ePXv27NAVXAAAAIAfkg4F8ejoaD3//PNyOp0qLi5WaWmpfvnLX3boBElJSUpKSpIkRUZGatWqVa22GThwoJYvX96qPTQ0VMuWLWvV7u/vryVLlnTo/AAAAMAPUYeCeGpqqpKTkzVx4kTV1dXp5Zdf1vTp0/Xss896uz4AAACgU+rQGvHKykr3chFfX1/ddtttLT5gCQAAAODEdPjDmk3X8JYOX9nEGOO1ogAAAIDOrkNLU2677TaNHTtWI0aMkM1mU2FhoR5++GFv1wYAAAB0Wh0K4uPGjVNYWJjef/99+fj46I477lD//v29XRsAAADQaXUoiEuHr2wycOBAb9YCAAAAnDFO7HvhAQAAAHgEQRwAAACwAEEcAPC9uerrrS7B4zpjnwD8sHR4jTgAAO3x6dpVb15/k9VleNQvV7xsdQkAOjlmxAEAAAALEMQBAAAACxDEAQAAAAsQxAEAAAALEMQBAAAACxDEAQAAAAsQxAEAAAALEMQBAPCgRmfn+yKgztgn4IeAL/QBAMCDuji66t93x1hdhkeFL15rdQlAp8SMOAAAAGABgjgAAABgAYI4AAAAYAGCOAAAAGABgjgAAABgAYI4AAAAYAGCOAAAAGABgjgAAABgAYI4AAAAYAGCOAAAAGABgjgAAABgAYI4AAAAYAGvBvGnn35a11xzjeLj47V06VJJUmFhoRITEzV69GjNmzfPve327duVlJSkmJgYpaWlqaGhQZK0Z88eJScnKzY2VnfffbdqamokSQcPHtSdd96puLg4JScnq7y83JtdAQAAADzKa0H8X//6l95//32tWrVKK1as0LJly/TZZ59p+vTpys7OVn5+voqKirRhwwZJUkpKimbMmKG1a9fKGKOcnBxJUmZmpiZMmKCCggKFhYUpOztbkjR//nxFRERozZo1Gj9+vGbNmuWtrgAAgJPQ2FBndQke1xn7BOvYvXXgSy+9VC+88ILsdrtKS0vlcrl08OBB9e3bV3369JEkJSYmqqCgQBdccIFqa2sVHh4uSUpKStKCBQs0fvx4bd68Wc8884y7/eabb1ZKSorWr1+vl156SZKUkJCgmTNnyul0yuFweKtLAADgBHSx++rLWYOtLsOjLkzbZnUJ6ES8FsQlyeFwaMGCBXr++ecVGxursrIyBQYGuu8PCgpSaWlpq/bAwECVlpaqsrJSfn5+stvtLdoltdjHbrfLz89P+/fvV+/evTtUW1FRUZvtQ4cOPam+ng62bt16QtszFs0Yi2addSxOdBwkxuJIjEUzxqIZYwEcm1eDuCTdf//9mjRpku666y7t3LlTNpvNfZ8xRjabTY2NjW22N/080tG3j9ynS5eOr7QJCwuTr6/vCfbm9NZZ/yCeDMaiGWNxGOPQjLFoxlg0YyyaMRboqLq6unYnfyUvrhH/6quvtH37dklS9+7dNXr0aH3wwQctPlRZXl6uoKAgBQcHt2ivqKhQUFCQAgICVF1dLZfL1WJ76fBsekVFhSSpoaFBNTU18vf391Z3AAAAAI/yWhDfvXu30tPTVV9fr/r6er399tu68cYbtWPHDu3atUsul0urV69WVFSUQkND5evr636rJy8vT1FRUXI4HIqIiFB+fr4kKTc3V1FRUZKk6Oho5ebmSpLy8/MVERHB+nAAAACcNry2NCU6Olrbtm3T2LFj5ePjo9GjRys+Pl4BAQGaPHmy6urqFB0drdjYWElSVlaW0tPTdejQIQ0aNEgTJ06UJGVkZCg1NVWLFy9WSEiI5s6dK0maMmWKUlNTFR8fr549eyorK8tbXQEAAAA8zqtrxCdPnqzJkye3aIuMjNSqVatabTtw4EAtX768VXtoaKiWLVvWqt3f319LlizxXLEAAADAKcQ3awIAAAAWIIgDAAAAFiCIAwAAABYgiAMAAAAWIIgDAAAAFiCIAwAAABYgiAMAAAAWIIgDAAAAFiCIAwAAABYgiAMAAAAWIIgDAAB4WaOrzuoSPK4z9ulUs1tdAAAAQGfXxcdXO1+62OoyPOqnyR9ZXcJpjxlxAAAAwAIEcQAAAMACBHEAAADAAgRxAAAAwAIEcQAAAMACBHEAAADAAgRxAAAAwAIEcQAAAMACBHEAAADAAgRxAAAAnDKm0Wl1CR53sn3iK+4BAABwyti6OLRr6+1Wl+FRfYc+f1L7MSMOAAAAWIAgDgAAAFiAIA4AAABYgCAOAAAAWIAgDgAAAFiAIA4AAABYgCAOAAAAWMCrQXzRokWKj49XfHy8nnzySUlSYWGhEhMTNXr0aM2bN8+97fbt25WUlKSYmBilpaWpoaFBkrRnzx4lJycrNjZWd999t2pqaiRJBw8e1J133qm4uDglJyervLzcm10BAAAAPMprQbywsFDvvvuuXn/9deXm5ur//u//tHr1ak2fPl3Z2dnKz89XUVGRNmzYIElKSUnRjBkztHbtWhljlJOTI0nKzMzUhAkTVFBQoLCwMGVnZ0uS5s+fr4iICK1Zs0bjx4/XrFmzvNUVAAAAwOO8FsQDAwOVmpqqrl27yuFw6Gc/+5l27typvn37qk+fPrLb7UpMTFRBQYFKSkpUW1ur8PBwSVJSUpIKCgrkdDq1efNmxcTEtGiXpPXr1ysxMVGSlJCQoI0bN8rp7HxfmQoAAIDOyWtB/MILL3QH6507d2rNmjWy2WwKDAx0bxMUFKTS0lKVlZW1aA8MDFRpaakqKyvl5+cnu93eol1Si33sdrv8/Py0f/9+b3UHAAAA8Ci7t0/w5Zdf6re//a0efvhh+fj4aOfOne77jDGy2WxqbGyUzWZr1d7080hH3z5yny5dOv66oqioqM32oUOHdvgYp5utW7ee0PaMRTPGollnHYsTHQeJsTgSY9GMsWjGWDRjLJoxFs28GsS3bt2q+++/X9OnT1d8fLz+9a9/tfhQZXl5uYKCghQcHNyivaKiQkFBQQoICFB1dbVcLpd8fHzc20uHZ9MrKioUHByshoYG1dTUyN/fv8O1hYWFydfX13OdPQ101gf+yWAsmjEWhzEOzRiLZoxFM8aiGWPRjLFo1tZY1NXVtTv5K3lxacrevXt17733KisrS/Hx8ZKkIUOGaMeOHdq1a5dcLpdWr16tqKgohYaGytfX1/1KIi8vT1FRUXI4HIqIiFB+fr4kKTc3V1FRUZKk6Oho5ebmSpLy8/MVEREhh8Phre4AAAAAHuW1GfHnnntOdXV1mjNnjrvtxhtv1Jw5czR58mTV1dUpOjpasbGxkqSsrCylp6fr0KFDGjRokCZOnChJysjIUGpqqhYvXqyQkBDNnTtXkjRlyhSlpqYqPj5ePXv2VFZWlre6AgAAAHic14J4enq60tPT27xv1apVrdoGDhyo5cuXt2oPDQ3VsmXLWrX7+/tryZIl379QAAAAwAJ8syYAAABgAYI4AAAAYAGCOAAAAGABgjgAAABgAYI4AAAAYAGCOAAAAGABgjgAAABgAYI4AAAAYAGCOAAAAGABgjgAAABgAYI4AAAAYAGCOAAAAGABgjgAAABgAYI4AAAAYAGCOAAAAGABgjgAAABgAYI4AAAAYAGCOAAAAGABgjgAAABgAYI4AAAAYAGCOAAAAGABgjgAAABgAYI4AAAAYAGCOAAAAGABgjgAAABgAYI4AAAAYAGCOAAAAGABgjgAAABgAYI4AAAAYAGCOAAAAGABrwbxQ4cOKSEhQbt375YkFRYWKjExUaNHj9a8efPc223fvl1JSUmKiYlRWlqaGhoaJEl79uxRcnKyYmNjdffdd6umpkaSdPDgQd15552Ki4tTcnKyysvLvdkNAAAAwOO8FsQ//vhj3XTTTdq5c6ckqba2VtOnT1d2drby8/NVVFSkDRs2SJJSUlI0Y8YMrV27VsYY5eTkSJIyMzM1YcIEFRQUKCwsTNnZ2ZKk+fPnKyIiQmvWrNH48eM1a9Ysb3UDAAAA8AqvBfGcnBxlZGQoKChIkrRt2zb17dtXffr0kd1uV2JiogoKClRSUqLa2lqFh4dLkpKSklRQUCCn06nNmzcrJiamRbskrV+/XomJiZKkhIQEbdy4UU6n01tdAQAAADzO7q0DHz1LXVZWpsDAQPftoKAglZaWtmoPDAxUaWmpKisr5efnJ7vd3qL96GPZ7Xb5+flp//796t27t7e6AwAAAHiU14L40RobG2Wz2dy3jTGy2Wzttjf9PNLRt4/cp0uXE5vcLyoqarN96NChJ3Sc08nWrVtPaHvGohlj0ayzjsWJjoPEWByJsWjGWDRjLJoxFs0Yi2anLIgHBwe3+FBleXm5goKCWrVXVFQoKChIAQEBqq6ulsvlko+Pj3t76fBsekVFhYKDg9XQ0KCamhr5+/ufUD1hYWHy9fX1TOdOE531gX8yGItmjMVhjEMzxqIZY9GMsWjGWDRjLJq1NRZ1dXXtTv5Kp/DyhUOGDNGOHTu0a9cuuVwurV69WlFRUQoNDZWvr6/7VUReXp6ioqLkcDgUERGh/Px8SVJubq6ioqIkSdHR0crNzZUk5efnKyIiQg6H41R1BQAAAPjeTtmMuK+vr+bMmaPJkyerrq5O0dHRio2NlSRlZWUpPT1dhw4d0qBBgzRx4kRJUkZGhlJTU7V48WKFhIRo7ty5kqQpU6YoNTVV8fHx6tmzp7Kysk5VNwAAAACP8HoQX7dunfvfkZGRWrVqVattBg4cqOXLl7dqDw0N1bJly1q1+/v7a8mSJZ4tFAAAADiF+GZNAAAAwAIEcQAAAMACBHEAAADAAgRxAAAAwAIEcQAAAMACBHEAAADAAgRxAAAAwAIEcQAAAMACBHEAAADAAgRxAAAAwAIEcQAAAMACBHEAAADAAgRxAAAAwAIEcQAAAMACBHEAAADAAgRxAAAAwAIEcQAAAMACBHEAAADAAgRxAAAAwAIEcQAAAMACBHEAAADAAgRxAAAAwAIEcQAAAMACBHEAAADAAgRxAAAAwAIEcQAAAMACBHEAAADAAgRxAAAAwAIEcQAAAMACBHEAAADAAqd1EP/73/+ua665RqNHj9ZLL71kdTkAAABAh9mtLuBklZaWat68eVq5cqW6du2qG2+8UcOGDdMFF1xgdWkAAADAcZ22M+KFhYW67LLL5O/vr7POOksxMTEqKCiwuiwAAACgQ07bGfGysjIFBga6bwcFBWnbtm3H3c8YI0mqr69vd5ug3r2+f4E/MHV1dSe1X0Dvsz1cifVOdizODuzp4Uqsd7JjcVYvPw9XYq2THQdJ6npu53pcfJ+x6HJO5/p78X3Gwvj5e7AS632fsWjs3rn+n/p9xsLlYCyauHRm/O1syptN+fNoNtPePT9wixcvVl1dnR544AFJUk5OjoqKijRz5sxj7lddXa0vvvjiVJQIAAAAqH///urZs/WLj9N2Rjw4OFhbtmxx3y4vL1dQUNBx9+vRo4f69+8vh8Mhm83mzRIBAABwBjPGyOl0qkePHm3ef9oG8csvv1wLFy7U/v371b17d73xxhv6/e9/f9z9unTp0uYrEgAAAMDTunXr1u59p20Q7927t6ZOnaqJEyfK6XRq3LhxGjx4sNVlAQAAAB1y2q4RBwAAAE5np+3lCwEAAIDTGUEcAAAAsABBHAAAALAAQRwAAACwAEEcAAAAsABB3ANGjRql3bt3W13Gcb3zzjtaunSpJeeurq7Wvffeq927d2vUqFGW1IAftqbHCFr65JNPlJaWppycHK1evdrqcjzi0KFDSkhIaPF38+GHH9bKlSvdt19//XVdccUVGjNmjMaMGaN58+ZJkqqqqjRp0iRde+21GjdunLZv3y5JKikp0cUXX+ze/o477nAf6w9/+IPi4+OVkJDQacbweFauXKnU1FSry/CqpufGmaYj/z8dM2bMKa4KJ+u0vY44TlxRUZFl566qqnL/DxNoC4+Rtl100UW66KKLlJqaqksvvdTqcr63jz/+WOnp6dq5c6ckqbS0VBkZGdq0aZMuu+wy93ZFRUVKTU1VQkJCi/2XLl2q/v3769lnn9W6des0c+ZMvfzyyyoqKlJiYqJmzpzZYvtNmzZp27ZtWrVqlSorKxUXF6errrpK3bt393pf4V1Nz40zTUf+Vubl5Z2iavB9ndFBvKGhQY899pi+/PJLVVRUaMCAAfrd736nBx54QOeff77+85//6LzzztNTTz0lf39/Rf6/9u48JqqrfeD4dxgQoSq7FrdqVbBiVFSqwLgA0roEZbMSKwSrsbQqrUpbVLRal1Zk1ApWTVdF/4DAoA7YNhYStYZFDFASAetabdwSSlUQgZn5/UGYwCu279tfmUF4Pv/InVwPzxnvPee5Z/F6exMYGEhJSQkvvPACSUlJDBw40FieTqcjMTGRoqIidDodoaGhREdHc+fOHeLi4qirq8PCwoKEhATGjRuHv78/hw8fZuDAgRQWFpKSkkJqaiqRkZGMHDmS4uJinjx5wrp161CpVMTHx2NtbU15eTm1tbW88847BAcH8/jxYxISEqiqqkKhULBkyRKCg4PRaDRkZWVRU1PDSy+9RElJCQD9+/cnLCzMpN/11q1buXfvHp9++in19fWsWrWKX3/9lT59+rBv3z4cHBw4c+YMe/fupampiYEDB7JlyxYcHBxMGqepFBYWcuDAAaysrIyjGra2tvz0008AzJw5k6qqKtRqNQDJyclYW1uzbNkyc4bdoVqukeXLl+Pv78+3336LQqHAw8ODDRs2PPP1wF1dYWEhe/bs4cqVKxQUFODi4sKUKVPMHdY/lp6ezscff8yHH34IgFarJSAgAHt7+zbnlZeXc/36dQ4ePIi7uzsbNmzAzs4OvV5PbW0tAI8fPza+sa68vJxLly4xb9487OzsWL9+Pe7u7uh0Op48eUJTUxOPHz+mR48epq3w/8Oz+qg1a9YwYsQIKioqcHJy4vPPP8fe3p5jx46xf/9+evXqxYABA7C1tTV3FTpUS7/p7+9PVlYWFhYWjBkz5qmHsa7mv+lP3d3dqaqqIj8/n507dwJgZ2eHWq3G0dHRzDXoOAcOHODEiRMolUp8fX1ZuHAhsbGx7d4vnUW3XppSUlKClZUVaWlpnDp1iocPH3L69GkuXbrEwoULycnJYdiwYaSkpABQXV2Np6cnWq2WOXPmsHXr1jblpaenA81TqhkZGeTm5lJcXExGRgbTp09Ho9EQGxvLhQsX/ja2R48ekZWVhVqtJj4+noaGBgBu3rxJWloahw4dIjExkfv375OcnIyDgwPZ2dkcOnSI5ORkKisrgebRpqysLFJSUoiIiCAiIsLkSThAQkICffv2Ze3atVRXV7N48WKys7Nxdnbm5MmTVFdXo1ar+frrrzl27BgqlYqkpCSTx2lKZWVlbN68mczMTI4ePYqjoyMajQZ3d3esra3Jz8/n0aNHAGRnZ3f5qcaWayQ2NpYDBw6QmpqKVqvFxsbGeA92V5aWlvj7+xMbG/tcJ+EA27ZtY+LEicbjpUuXMn/+/KfOc3Fx4d133+XEiRO4uroak6u33nqL/Px8VCoVCQkJxMbGAmBtbc3cuXPJyspiyZIlLF++nIaGBlQqFYMGDWLq1KnMnj2bZcuWPTej4c/qoyorK41taJ8+fdBqtdy9e5ekpCSOwe0pyQAACVhJREFUHj1KWlqa8WGlq9PpdBw8eJDMzEw0Gg2NjY3cvXvX3GF1qL/rT1v74osv2LRpExqNBh8fHy5evGimqDve6dOnycvLIzMzk6ysLG7cuMHZs2fbvV86k249Iu7l5YW9vT1Hjx7l6tWrXL9+nbq6OoYMGcKkSZMACA4OJi4uDmhu6IODgwEICQlh165dbcrLz8+noqKCgoICAOrq6qiqqsLb25uVK1dSUVHBtGnTWLRo0d/G9sYbbwDwyiuv4OLiQlVVFQChoaFYWVnx4osvMn78eC5cuEBBQQHbt28HwNHRkYCAAIqKiujVqxejRo3C0rJz/TP37duXMWPGADB8+HD++OMPysrKuH37NlFRUQDo9Xrs7OzMGWaHc3Nzw9XVFQAHBwe8vb2B5hmLBw8eMG3aNE6dOsWgQYMYNGgQ/fr1M2e4JnP+/Hn8/PyMsyELFixg7dq1Zo5KmNq+ffuMPy9dupTAwEAAtmzZwptvvklUVBQlJSWsWrWKnJwcVq5caTx/2rRpqNVqrl69SllZGUqlkp9//pmamhqioqIYO3Ys48aNM3md/lfP6qOcnJwYNWoUACNGjODPP/+kpKQET09PnJ2dAQgKCjL2RV2ZUqnE09OT8PBwAgICWLx4cbdpK6H9/rS1gIAAVqxYwYwZMwgICMDX19ccYZpEQUEBc+bMMT5oh4WFcezYsXbvl86kW4+I5+bmEhcXR8+ePQkNDcXLy4v+/fu3SVwNBgNKpRIACwsLFAoF0JwotnzeQqfT8cEHH3D8+HGOHz9OWloa4eHhTJgwgZycHFQqFSdPniQmJqZN+dA8Bdla67L1er0xpvY+bymjdZk6nQ7AOG3bmbT+fhUKhTHe8ePHG7+7jIwM9u7da8YoO56VlVWb4/+8nsLCwsjOzkar1RIaGmrK0MxKr9e3OTYYDE/dH6Jre/jwId99953xuHU7nJuba5zV8/T0xMnJiStXrpCamtomCTEYDFhaWpKbm8vcuXOxsrLCxcWF6dOnU1xcbNL6/FPP6qOsra2N57S0oS1/tuhsAzAdqWXU12AwsHTpUoqKiswdksm015+2Fh0dTWpqKoMHD2bnzp3s37/f1CGazH/2HdCcW7V3v3Qm3ToRz8/PZ9asWYSFhdGnTx8KCwvR6XRcu3bNuBEiMzOTqVOnAs3rEfPy8oDmHektn7eYPHky6enpNDY2Ultby8KFCyktLSUxMZETJ04QEhLCxo0bjVNDDg4OXL58GWhucFtrmV4qLy/nwYMHuLm5AfD9999jMBj4/fff+eWXX5gwYQKTJ08mIyMDaF4+k5ub2+6mLqVSabaExtLS8i9/99ixYyktLeXatWtAc8OamJhoqvA6pYkTJ3Lnzh0KCwuZMWOGucPpcC3XyKuvvkpeXh41NTVA85Kvlhmq7kypVBofsLs6W1tbvvrqK8rKygA4cuSIcUR85MiRxr0U169f5969ewwdOpTz588b28GioiL0ej0vv/xym/Pr6uooKChg9OjRZqjV/+5ZfVR7JkyYQGlpKXfv3kWv1z+1RKGrqq6uZvbs2bi5ufHee+/h6+trnEHuqv6uP21t/vz51NbWEh0dTXR0dJdemjJ58mRycnKor6+nqamJzMzMNhvAO6vu88jcjvnz5xMXF0dOTg5WVlaMHz+ewsJC7Ozs2Lt3L7/99hvu7u5t1oL/8MMP7N69m759+7Jjx4425UVERHDjxg1CQkJoamoiNDSUSZMmMXjwYNasWYNGo0GpVBr/XmxsLFu2bCElJQWVStWmrJs3bxISEgLA7t27jaNB9fX1hIWF0dDQwCeffIKDgwPLly9n06ZNBAUFodPpiImJwcPD46nGyMvLi48++ghnZ2ciIyP/9e/zrzg5OdG/f/9nLjFwcXFh+/btvP/+++j1evr162fcYNKdBQYGUlNT81xtMPunWq6Rbdu28fbbbxMZGUljYyMeHh5s3rzZ3OGZnY+PD7t27aJ3797MnDnT3OF0KKVSyZ49e9i0aRP19fUMGTLE+GD+2WefsXHjRr788kt69OjBjh076N27N+vXryc+Pp7jx49jbW2NWq3GwsKCmJgYNm/ezKxZs1AqlYSHhz8XnTM8u49qj7OzMwkJCURHR2NjY8Pw4cNNHK15tCzHDA8Px8bGhqFDh5plH5Qp/V1/2trq1auJj4/H0tISW1vbp/a2dSV+fn5UVFQQFhZGU1MTKpUKPz8/Dh8+bO7Q/pLC0NnG6M3s1q1bREVFGUe+W2vZhdzRIiMjWbFixVOjgC3/fVl3WqbQXRkMBhobG1m8eDHr1q3Dw8PD3CEJIYQQ4l/WrZemCNFZ3b9/H19fX8aOHStJuBBCCNFFyYi4EEIIIYQQZiAj4kIIIYQQQpiBJOJCCCGEEEKYgSTiQgghhBBCmIEk4kIIIf411dXVuLu7mzsMIYR4LkgiLoQQQgghhBl06xf6CCFEdzVv3jzi4+Px9vYmOzubtWvXcv78eXr27Mn69etxc3OjvLycyspKFAoFU6ZMYfXq1VhaWjJ69GgCAgKorKwkKSmJ27dvs3v3bmxsbJ6bt1YKIURnICPiQgjRDQUGBnLmzBkAzp49i52dHcXFxRgMBk6fPs2ZM2ewt7dHq9WSmZlJVVUV33zzDQCNjY34+fnx448/4urqyrp160hOTkaj0TBgwABzVksIIZ4rkogLIUQ31JKIGwwGiouLiY6O5ty5c5SWljJ48GAuXrzIokWLUCgU9OjRg4iICGPiDjBx4kQALly4gJubm/GV6gsWLDBLfYQQ4nkkibgQQnRD7u7uNDY2kpuby5AhQ/Dz8+PcuXPk5eXx+uuvo9frUSgUxvP1ej1NTU3GY1tbW+PPrd8LZ2kpKx6FEOK/JYm4EEJ0UzNmzECtVuPr68uwYcN49OgRWq2W1157DZVKxZEjRzAYDDQ0NJCeno6Pj89TZXh5eXH58mUqKysB0Gg0pq6GEEI8tyQRF0KIbiowMJCrV68aE2wfHx9cXFxwdXUlISGB6upqgoKCCAoKYujQocTExDxVhqOjI0lJScTFxRESEsKtW7dMXQ0hhHhuKQyt5xSFEEIIIYQQJiEj4kIIIYQQQpiBJOJCCCGEEEKYgSTiQgghhBBCmIEk4kIIIYQQQpiBJOJCCCGEEEKYgSTiQgghhBBCmIEk4kIIIYQQQpiBJOJCCCGEEEKYwf8BA36b+myB7WoAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 864x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "top10_bagofwords(inbound, 'most_common_before' ,'Top 10 Most Common Words in My Inbound Data Before Preprocessing')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We see that after preprocessing, we will remove the handles which is what applesupport and 115858 is in the visualization above. For now, I briefly visualized the top 10 most common tokens to show that my data as it currently is quite dirty, they will definitely skew and ruin my results if I don't do preprocessing!"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Emoji Analysis\n",
    "You can leave emojis in and have model learn embeddings for the emojis. Emoji are used in very specific instances, if they use emojis it will be in a useful sense. If the emojis are used for inside jokes, then it's different story, but that is likely not the case for this dataset because these are customer service tweets.\n",
    "\n",
    "There's actually an embedding called emoji2vec, but that's something I can explore down the line. Right now, I think doc2vec won't be able to handle these emojis in the best way and it might skew my results if I leave them in.\n",
    "\n",
    "Extracted from [this kernel](https://www.kaggle.com/psbots/customer-support-meets-spacy-universe)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import en_core_web_lg\n",
    "\n",
    "# This works, I couldn't load the en_core_web_lg\n",
    "nlp = spacy.load('en')\n",
    "emoji = Emoji(nlp, merge_spans=False)\n",
    "nlp.add_pipe(emoji, first=True)\n",
    "print(nlp.pipe_names)\n",
    "\n",
    "# This is a list that stores all the emojis\n",
    "# Make sure to have your processed_inbound ready. I used an intermediate version\n",
    "# of my processed data. Or you can run this after you've run my preprocessing\n",
    "# pipeline below.\n",
    "\n",
    "# processed_inbound.progress_apply(\" \".join)\n",
    "emojis = []\n",
    "for doc in tqdm(nlp.pipe(inbound, batch_size=512)):\n",
    "    if doc._.has_emoji:\n",
    "        for e in doc._.emoji:\n",
    "            emojis.extend(e[0])\n",
    "\n",
    "# Let's count them\n",
    "emoji_count = Counter(emojis)\n",
    "emoji_count.most_common(30)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's visualize the emojis as well and export it as png"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from bokeh.io import export_png\n",
    "\n",
    "# Let's store it in a list, so we can remove it from the Pipeline above\n",
    "unique_emojis = [i[0] for i in emoji_count]\n",
    "\n",
    "# Also storing the counts in a list. Skipping irrelevant, unintepretable ones.\n",
    "plotted_emojis = [i[0] for i in emoji_count.most_common(30)]\n",
    "plotted_emojis = plotted_emojis[:3] + plotted_emojis[5:]\n",
    "counts = [i[1] for i in emoji_count.most_common(30)]\n",
    "counts = counts[:3] + counts[5:]\n",
    "\n",
    "# Using Bokeh because pyplot doesn't show emojis\n",
    "# Run this cell to view the html visualization\n",
    "\n",
    "output_file(\"visualizations/emojis.html\")\n",
    "\n",
    "\n",
    "p = figure(x_range=plotted_emojis, plot_height=600, plot_width = 1300, title=\"Emoji Counts\",\n",
    "           toolbar_location=None, tools=\"\")\n",
    "p.vbar(x=plotted_emojis, top=counts, width=0.9, color = \"#FBD043\")\n",
    "p.xgrid.grid_line_color = None\n",
    "p.y_range.start = 0\n",
    "\n",
    "export_png(p, filename=\"visualizations/emojis.png\")\n",
    "\n",
    "show(p)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "filename = \"visualizations/emojis.html\"\n",
    "IFrame(src=filename, width = 800, height=700)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Here we mostly see negative emojis! Meaning that most of the sentiment of the customers are complaints or inquiries!\n",
    "\n",
    "For now, I will remove the emojis, however in the future it might be interesting to do stuff with these emojis! Or, as an alternative I might leave only the most informative and common ones, maybe like the top 10 that clearly expresses a sentiment - which is something that will be related to an intent."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Building my own tokenizer \n",
    "\n",
    "### Variable storage\n",
    "punct_base = ['?']\n",
    "\n",
    "# Punctuations I want to remove, including the empty token\n",
    "puncts = ['\\u200d','?', '....','..','...','','@','#', ',', '.', '\"', ':', ')', '(', '-', '!', '|', ';', \"'\", '$', '&', '/', '[', ']', '>', '%', '=', '*', '+', '\\\\', \n",
    "    '•', '~', '£', '·', '_', '{', '}', '©', '^', '®', '`',  '<', '→', '°', '€', '™', '›',  '♥', '←', '×', '§', '″', '′', 'Â', '█', \n",
    "    '½', 'à', '…', '“', '★', '”', '–', '●', 'â', '►', '−', '¢', '²', '¬', '░', '¶', '↑', '±', '¿', '▾', '═', '¦', '║', '―', '¥', '▓', \n",
    "    '—', '‹', '─', '▒', '：', '¼', '⊕', '▼', '▪', '†', '■', '’', '▀', '¨', '▄', '♫', '☆', 'é', '¯', '♦', '¤', '▲', 'è', '¸', '¾', \n",
    "    'Ã', '⋅', '‘', '∞', '∙', '）', '↓', '、', '│', '（', '»', '，', '♪', '╩', '╚', '³', '・', '╦', '╣', '╔', '╗', '▬', '❤', 'ï', 'Ø', \n",
    "    '¹', '≤', '‡', '√', '!','🅰','🅱']\n",
    "\n",
    "# Using NLTK's stop words corpus\n",
    "stopwords.words('english');\n",
    "stop_words = set(stopwords.words('english')) \n",
    "\n",
    "# Found a dictionary of common contractions and colloquial language\n",
    "contraction_colloq_dict = {\"btw\": \"by the way\", \"ain't\": \"is not\", \"aren't\": \"are not\",\"can't\": \"cannot\", \"'cause\": \"because\", \"could've\": \"could have\", \"couldn't\": \"could not\", \"didn't\": \"did not\",  \"doesn't\": \"does not\", \"don't\": \"do not\", \"hadn't\": \"had not\", \"hasn't\": \"has not\", \"haven't\": \"have not\", \"he'd\": \"he would\",\"he'll\": \"he will\", \"he's\": \"he is\", \"how'd\": \"how did\", \"how'd'y\": \"how do you\", \"how'll\": \"how will\", \"how's\": \"how is\",  \"I'd\": \"I would\", \"I'd've\": \"I would have\", \"I'll\": \"I will\", \"I'll've\": \"I will have\",\"I'm\": \"I am\", \"I've\": \"I have\", \"i'd\": \"i would\", \"i'd've\": \"i would have\", \"i'll\": \"i will\",  \"i'll've\": \"i will have\",\"i'm\": \"i am\", \"i've\": \"i have\", \"isn't\": \"is not\", \"it'd\": \"it would\", \"it'd've\": \"it would have\", \"it'll\": \"it will\", \"it'll've\": \"it will have\",\"it's\": \"it is\", \"let's\": \"let us\", \"ma'am\": \"madam\", \"mayn't\": \"may not\", \"might've\": \"might have\",\"mightn't\": \"might not\",\"mightn't've\": \"might not have\", \"must've\": \"must have\", \"mustn't\": \"must not\", \"mustn't've\": \"must not have\", \"needn't\": \"need not\", \"needn't've\": \"need not have\",\"o'clock\": \"of the clock\", \"oughtn't\": \"ought not\", \"oughtn't've\": \"ought not have\", \"shan't\": \"shall not\", \"sha'n't\": \"shall not\", \"shan't've\": \"shall not have\", \"she'd\": \"she would\", \"she'd've\": \"she would have\", \"she'll\": \"she will\", \"she'll've\": \"she will have\", \"she's\": \"she is\", \"should've\": \"should have\", \"shouldn't\": \"should not\", \"shouldn't've\": \"should not have\", \"so've\": \"so have\",\"so's\": \"so as\", \"this's\": \"this is\",\"that'd\": \"that would\", \"that'd've\": \"that would have\", \"that's\": \"that is\", \"there'd\": \"there would\", \"there'd've\": \"there would have\", \"there's\": \"there is\", \"here's\": \"here is\",\"they'd\": \"they would\", \"they'd've\": \"they would have\", \"they'll\": \"they will\", \"they'll've\": \"they will have\", \"they're\": \"they are\", \"they've\": \"they have\", \"to've\": \"to have\", \"wasn't\": \"was not\", \"we'd\": \"we would\", \"we'd've\": \"we would have\", \"we'll\": \"we will\", \"we'll've\": \"we will have\", \"we're\": \"we are\", \"we've\": \"we have\", \"weren't\": \"were not\", \"what'll\": \"what will\", \"what'll've\": \"what will have\", \"what're\": \"what are\",  \"what's\": \"what is\", \"what've\": \"what have\", \"when's\": \"when is\", \"when've\": \"when have\", \"where'd\": \"where did\", \"where's\": \"where is\", \"where've\": \"where have\", \"who'll\": \"who will\", \"who'll've\": \"who will have\", \"who's\": \"who is\", \"who've\": \"who have\", \"why's\": \"why is\", \"why've\": \"why have\", \"will've\": \"will have\", \"won't\": \"will not\", \"won't've\": \"will not have\", \"would've\": \"would have\", \"wouldn't\": \"would not\", \"wouldn't've\": \"would not have\", \"y'all\": \"you all\", \"y'all'd\": \"you all would\",\"y'all'd've\": \"you all would have\",\"y'all're\": \"you all are\",\"y'all've\": \"you all have\",\"you'd\": \"you would\", \"you'd've\": \"you would have\"}\n",
    "\n",
    "# Initializing the lemmatizer\n",
    "lemmatizer = nltk.stem.WordNetLemmatizer()\n",
    "\n",
    "import emoji # For emoji removal step\n",
    "\n",
    "import en_core_web_sm\n",
    "try:\n",
    "    # Initializing spaCy objects (Don't run this more than once)\n",
    "    nlp_cld = spacy.load('en',disable_pipes=[\"tagger\",\"ner\"])\n",
    "    language_detector = spacy_cld.LanguageDetector()\n",
    "    nlp_cld.add_pipe(language_detector)\n",
    "except ValueError as e:\n",
    "    print(\"The spaCy function was run more than once, but that's okay because it means it was initialized!\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## My Entire Preliminary Text Preprocessing Pipeline\n",
    "\n",
    "I wanted to all the preprocessing to be an iterative process, so I compiled it all in one function so that it all runs at the same time and so there is an enhanced sense of organization. At every preprocessing step, I visualize the lengths of each tokens at the data. In general, things like removing stop-words will shift the distribution to the left because we have fewer and fewer tokens at every preprocessing step. I also provide a peek to the head of the data at each step so that it clearly shows what processing is being done at each step."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "# My preprocessing functions (defining them here so that I could access them from anywhere in the notebook)\n",
    "\n",
    "def visualize_lengths(data,title):\n",
    "    '''Visualizing lengths of tokens in each tweet'''\n",
    "    lengths = [len(i) for i in data]\n",
    "    plt.figure(figsize=(13,6))\n",
    "    plt.hist(lengths, bins = 40)\n",
    "    plt.title(title)\n",
    "    plt.show()\n",
    "    \n",
    "def remove_from_list(x, stuff_to_remove) -> list:\n",
    "        ''' Making a function to remove a list of items from a list'''\n",
    "        for item in stuff_to_remove:\n",
    "            # Making sure to iterate through the entire token\n",
    "            for i,token in enumerate(x):\n",
    "                if item == token:\n",
    "                    del x[i]\n",
    "        return x\n",
    "    \n",
    "def remove_links(doc):\n",
    "    return [re.sub(r'^https?:\\/\\/.*[\\r\\n]*', '', token, flags=re.MULTILINE) for token in doc]\n",
    "\n",
    "def correct_spellings(x):\n",
    "    ''' Takes as input a list and outputs a list of the corrected spelling'''\n",
    "    corrected_text = []\n",
    "    for word in x:\n",
    "        if word in x:\n",
    "            corrected_text.append(spell.correction(word))\n",
    "        else:\n",
    "            corrected_text.append(word)\n",
    "    return corrected_text\n",
    "\n",
    "def replace_from_dict(x,dic):\n",
    "    ''' Making a function to replace all the items in a list based on a dictionary. I made sure to use the insert\n",
    "            method to insert the longer-gram replacement as distinct items in the list at that specific index'''\n",
    "    replaced_counter = 0\n",
    "    for item in dic.items():\n",
    "        for i, e in enumerate(x):\n",
    "            if e == item[0]:\n",
    "                replaced_counter+=1\n",
    "                    # Inserting the expanded tokens in a way that preserves the order\n",
    "                del x[i]\n",
    "                for ix, token in enumerate(item[1].split()):\n",
    "                    x.insert(i+ix,token)\n",
    "#         print(f\"Amount of words replaced: {replaced_counter}\")\n",
    "    return x    \n",
    "\n",
    "def only_english(x):\n",
    "    ''' Making a function that only accepts English by appending True if it is English and False otherwise \n",
    "            into a mask. Returns a mask'''    \n",
    "    mask = []\n",
    "    x = x.apply(\" \".join)\n",
    "    try:\n",
    "        for i,doc in tqdm(enumerate(nlp_cld.pipe(x, batch_size=512))):\n",
    "            if 'en' not in doc._.languages or len(doc._.languages) != 1:\n",
    "                mask.append(False)\n",
    "            else:\n",
    "                mask.append(True)\n",
    "    except Exception as e:\n",
    "        print(f\"Exception:{e}\")\n",
    "    return mask\n",
    "\n",
    "def get_wordnet_pos(word):\n",
    "    \"\"\"Map POS tag to first character lemmatize() accepts\"\"\"\n",
    "    tag = nltk.pos_tag([word])[0][1][0].upper()\n",
    "    tag_dict = {\"J\": wordnet.ADJ,\n",
    "                    \"N\": wordnet.NOUN,\n",
    "                    \"V\": wordnet.VERB,\n",
    "                    \"R\": wordnet.ADV}\n",
    "    return tag_dict.get(tag, wordnet.NOUN)\n",
    "    \n",
    "def lemmatize_list(x):\n",
    "    ''' This lemmatizer function should work on a single list of tokenized data'''\n",
    "#         # Turning list into a string\n",
    "    x = \" \".join(x)\n",
    "    # Returning a list again\n",
    "    return [lemmatizer.lemmatize(w, get_wordnet_pos(w)) for w in nltk.word_tokenize(x)]\n",
    "\n",
    "def extract_emojis(s):\n",
    "        return [c for c in s if c not in emoji.UNICODE_EMOJI]\n",
    "\n",
    "def limit_length(x, max_tokens, min_tokens):\n",
    "    ''' Inputs a list and drops it out of the document if \n",
    "        the document has more than the max and less than the min'''\n",
    "    output = x\n",
    "    if len(x) > max_tokens:\n",
    "        output = np.nan\n",
    "    if len(x) <= min_tokens:\n",
    "        output = np.nan\n",
    "    return output\n",
    "\n",
    "def clean_numbers(x):\n",
    "    for i,j in enumerate(x):\n",
    "        if bool(re.search(r'\\d', j)):\n",
    "            del x[i]\n",
    "    return x\n",
    "\n",
    "def validate(func, locals):\n",
    "    ''' Validating a function below to accept correct input'''\n",
    "    for var, test in func.__annotations__.items():\n",
    "        value = locals[var]\n",
    "        try: \n",
    "            pr=test.__name__+': '+test.__docstring__\n",
    "        except AttributeError:\n",
    "            pr=test.__name__   \n",
    "        msg = '{}=={}; Test: {}'.format(var, value, pr)\n",
    "        assert test(value), msg\n",
    "\n",
    "# End to end tokenizer function\n",
    "\n",
    "def my_tokenizer(data: lambda _data: isinstance(_data, pd.Series)) -> 'Cleaned Pandas Series':\n",
    "    ''' I am making my own end-to-end tokenizer function for preprocessing that accepts\n",
    "    a Pandas Series as input and outputs a preprocessed Pandas Series'''\n",
    "    \n",
    "    # Making sure input is a series (these are two ways of doing the same thing)\n",
    "    assert isinstance(data,pd.Series), 'Input must be a Pandas Series'\n",
    "#     validate(my_tokenizer, locals())\n",
    "    \n",
    "    # 1. Converting all to lower case\n",
    "    data = data.str.lower()\n",
    "    \n",
    "    print(f'1. Original shape of data is {data.shape}')\n",
    "    \n",
    "    \n",
    "    # 2. Tokenizing with NLTK's TweetTokenizer. This limits repeated characters to \n",
    "    # three with the reduce lens paramater and strips all the @'s. It also\n",
    "    # splits it into 1-gram tokens\n",
    "    tknzr = TweetTokenizer(strip_handles = True, reduce_len = True)\n",
    "    # Using progress apply to show the progress bar\n",
    "    data = data.progress_apply(tknzr.tokenize)\n",
    "    \n",
    "    print(f'2. Tokenized, removed handles, and reduced the length of repeated characters.\\\n",
    "          \\n Shape is still {data.shape}. \\n \\n Peek: \\n {data.head()}')\n",
    "    visualize_lengths(data, 'Length of Tokens after Step 2')\n",
    "    \n",
    "    # 3. Removing the punctuation\n",
    "    \n",
    "    data = data.progress_apply(remove_from_list, stuff_to_remove = puncts)\n",
    "    print(f'3. Removed empty tokens and punctuation. Shape is still {data.shape}. \\n \\n Peek: \\n {data.head()}')\n",
    "    visualize_lengths(data, 'Length of Tokens after Step 9')\n",
    "    \n",
    "    # 3.1. Removing links\n",
    "    \n",
    "    data = data.progress_apply(remove_links)\n",
    "    \n",
    "    print(f'3. Removed the links. Shape is still {data.shape}. \\n \\n Peek: \\n {data.head()}')\n",
    "    visualize_lengths(data, 'Length of Tokens after Step 3')\n",
    "\n",
    "    # 4. Checking for and correcting mispellings\n",
    "    spell = SpellChecker()\n",
    "    \n",
    "    # Skipping this step first because it takes way too long\n",
    "#     data = data.progress_apply(correct_spellings)\n",
    "    \n",
    "    print(f'4. Applied automispelling corrections. Shape is still {data.shape}. \\n \\n Peek: \\n {data.head()}')\n",
    "    \n",
    "    \n",
    "    # 5. Removing the stop words, utilizing the same remove_from_list function defined above\n",
    "    \n",
    "    data = data.progress_apply(remove_from_list, stuff_to_remove = stop_words)\n",
    "    \n",
    "    print(f'5. Removed the stop words. Shape is still {data.shape}. \\n \\n Peek: \\n {data.head()}')\n",
    "    visualize_lengths(data, 'Length of Tokens after Step 4')\n",
    "    \n",
    "    # 6. Expanding contractions and colloquial language\n",
    "    \n",
    "    data = data.progress_apply(replace_from_dict, dic = contraction_colloq_dict)\n",
    "    print(f'6. Expanded contractions into extra tokens. Shape is still {data.shape}. \\n \\n Peek: \\n {data.head()}')\n",
    "    visualize_lengths(data, 'Length of Tokens after Step 5')\n",
    "    \n",
    "    # 7. Removing non-english Tweets with spaCy\n",
    "\n",
    "    data = data[only_english(data)]\n",
    "    print(f'7. Remove all non-english Tweets. Shape is now {data.shape}. Clearly less than before! \\n \\n Peek: \\n {data.head()}')\n",
    "    \n",
    "    # 8. Lemmatization\n",
    "\n",
    "    data = data.progress_apply(lemmatize_list)\n",
    "    print(f'8. Lemmatized the tokens. Shape is still {data.shape}. \\n \\n Peek: \\n {data.head()}')\n",
    "    visualize_lengths(data, 'Length of Tokens after Step 8')   \n",
    "   \n",
    "    # Removing again to make sure I get everything\n",
    "    data = data.progress_apply(remove_from_list, stuff_to_remove = puncts)\n",
    "\n",
    "    # 9.1. Removing emojis -- (UPDATE) in a way that preserves Series indexes\n",
    "    \n",
    "    unique_emojis = [i[0] for i in emoji.UNICODE_EMOJI]\n",
    "    data = data.apply(remove_from_list, stuff_to_remove = unique_emojis)\n",
    "    \n",
    "    # 9.2. Removing numbers -- (UPDATE) also in a way that preserves Series indexes\n",
    "    # Using nested list comprehension\n",
    "    data = data.progress_apply(clean_numbers)\n",
    "    \n",
    "    print(f'9. Removed emojis and numbers. Shape is still {data.shape}. \\n \\n Peek: \\n {data.head()}')\n",
    "    visualize_lengths(data, 'Length of Tokens after Step 9')   \n",
    "    \n",
    "    # 10. Limiting length of Tweet\n",
    "    max_tokens = 50\n",
    "    min_tokens = 5\n",
    "    data = data.progress_apply(limit_length, min_tokens = min_tokens, max_tokens = max_tokens)\n",
    "    # Dropping all nan values, which are the token limits that didn't meet the thresholding requirements\n",
    "    data = data.dropna()\n",
    "    print(f'10. Limited each tweet to a max of {max_tokens} tokens and a min of {min_tokens} tokens. Shape is now {data.shape}. \\n \\n Peek: \\n {data.head()}')\n",
    "    visualize_lengths(data, 'Length of Tokens after Step 10')\n",
    "    \n",
    "    return data\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1. Original shape of data is (106648,)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "86ebb316fd2e41aa91ceb8a2c5f751ea",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=106648.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "2. Tokenized, removed handles, and reduced the length of repeated characters.          \n",
      " Shape is still (106648,). \n",
      " \n",
      " Peek: \n",
      " 0    [the, newest, update, ., i️, made, sure, to, d...\n",
      "1                            [https://t.co/nv0yucs0lb]\n",
      "2    [tried, resetting, my, settings, .., restartin...\n",
      "3    [this, is, what, it, looks, like, https://t.co...\n",
      "4    [i️, have, an, iphone, 7, plus, and, yes, i️, do]\n",
      "Name: inbound_text, dtype: object\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 936x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "b1c10bfe66774955a94f09e9e759cc62",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=106648.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "3. Removed empty tokens and punctuation. Shape is still (106648,). \n",
      " \n",
      " Peek: \n",
      " 0    [the, newest, update, i️, made, sure, to, down...\n",
      "1                            [https://t.co/nv0yucs0lb]\n",
      "2    [tried, resetting, my, settings, restarting, m...\n",
      "3    [this, is, what, it, looks, like, https://t.co...\n",
      "4    [i️, have, an, iphone, 7, plus, and, yes, i️, do]\n",
      "Name: inbound_text, dtype: object\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 936x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "4757449691554c48b670b5c31224e9ce",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=106648.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "3. Removed the links. Shape is still (106648,). \n",
      " \n",
      " Peek: \n",
      " 0    [the, newest, update, i️, made, sure, to, down...\n",
      "1                                                   []\n",
      "2    [tried, resetting, my, settings, restarting, m...\n",
      "3                  [this, is, what, it, looks, like, ]\n",
      "4    [i️, have, an, iphone, 7, plus, and, yes, i️, do]\n",
      "Name: inbound_text, dtype: object\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAwsAAAF2CAYAAAAhhQSMAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3df1jU5Z7/8dcIA5p4jksx4nI8Xp2OLq2UttAW5UK5XgICS412SkhqOx5PbsfUbS38EVxWlnaR2J4Odk65nevi2FWEBcqFY22m/aBMuSpjta1OQqWGkKb8kGEYPvtH3+YrcvuDmbFBfT7+gbnn/nzmfc/c4rzm/nzmY7MsyxIAAAAAnGBQqAsAAAAAMDARFgAAAAAYERYAAAAAGBEWAAAAABgRFgAAAAAYERYAAAAAGBEWAOAMff3117rqqqt+lMfatWuXCgsLJUnbt29XVlZWQPs7cOCAsrKylJOTow8++MDX/qc//Uk5OTnKycnRVVddpUmTJvluf/nll8Z9FRQUaO3atQHVE0xFRUWaNGmSSkpK9NJLL2ndunUB73Pr1q269dZb9S//8i/KzMzUvHnz9M0330iSWltblZ+fH/BjHO8vf/mLMjMzlZWVpTlz5ujbb78N6v4BwF/hoS4AANDX559/rqampqDtb/v27brkkkv05z//uVf77NmzNXv2bEnSzJkzlZeXp/T09KA97o/hxRdf1NatWxUbG6uCggKNGTMmoP01NTXpgQce0Msvv6y4uDhJ0po1azR//ny98MILOnLkiD7++ONglC5Jqq+v13/913+pqqpKw4YN08qVK/Xkk0/qoYceCtpjAIC/CAsAEARdXV0qLi7Wjh075PV69fd///daunSpoqKiNGnSJN1888169913deDAAeXk5Gj+/PmSvv9kv6KiQkOHDlVSUpJef/11rVu3Tv/5n/+p1tZWLVq0SDfddJM6Ojq0YMECffHFF3K73XrkkUeUlJTUp44XX3xRZWVlGjRokC655BI9+OCDampq0urVq9Xa2qqZM2eqrKzsjMZk2tell17aq8+jjz6q//3f/1Vpaansdnu/n4P29nYtWrRIjY2NGjRokMaNG6eHHnpIgwb1Xvh+44039Mc//lFdXV06dOiQbrrpJs2fP1+5ubmyLEu/+c1vlJ6eri1btuidd97R4MGDlZeXpzVr1ujVV19VT0+P4uLiVFRUpBEjRmjmzJn66U9/qi+++EIzZszQzJkzfY91+PBheTwedXR0+NruuOMOxcfHS5IWLVqkzs5O5eTk6OWXX1ZDQ4OWL1+u7777Tl6vVzNnztT06dO1fft2FRcX62//9m/1xRdfaPDgwVqxYoUuu+yyXmNLSEjQ5s2bZbfb5Xa71dTUpJ/97Gdn9BoBwFlnAQDOyFdffWVNmDDBeN/vf/97a8WKFVZPT49lWZb1xBNPWEVFRZZlWdaNN95orVixwrIsy/rmm2+sK664wvryyy+tN99800pLS7OOHDli9fT0WIsWLbJuvPFGy7Isa/369dbs2bMty7Ks9957z7r88sutDz/80LIsy3ruuees/Pz8PjXU1tZakydPtr799lvfPjIyMqyenp5e+zuZ22+/3dq0adNp9/XAAw9YzzzzjLVs2TLrnnvusdxut9/PwSuvvGLdddddlmVZVnd3t7VkyRKroaGhV109PT3W7bffbu3du9e3/eWXX+6rbezYsb7fH3jgAevZZ5+1LMuyXnnlFWv+/PmWx+OxLMuyXnjhBWvWrFm+sS5atOikz8Vjjz1mjRs3zsrIyLCWLFliVVdX+/Zz/DzweDzW1KlTrfr6esuyLOvo0aNWRkaG9cEHH1jvvfeeFR8fb+3YscOyLMt6/vnnrZtvvvmkj/naa69Z//iP/2hNnDjRN1YACDVWFgAgCLZu3arW1lbV1tZKkjwejy6++GLf/f/8z/8sSRoxYoQuvvhiHTlyRNu2bVN6erp+8pOfSJLy8vL03nvvGfc/atQojR8/XpIUHx+v9evX9+nz1ltvaerUqYqOjpYkOZ1OLV++XF9//XW/x3O6ff35z3/Wt99+q8rKSkVERPj9HCQmJqqkpEQzZ87UddddpzvuuEOjR4/uVYvNZtPTTz+trVu3qrq6Wn/9619lWZaOHTt2yjG88cYb+vjjjzVt2jRJUk9PT69tTCszPygoKNBvf/tbvf/++9qxY4cef/xxlZWV9TkfoqGhQV9++aUWL17sa+vs7NTu3bt12WWXKT4+3vc406ZN00MPPaTDhw/rb/7mb/o85uTJkzV58mSVl5fr17/+tV577bU+KywA8GMjLABAEPT09Gjx4sVKTU2VJLW3t8vtdvvuj4yM9P1us9lkWZbCw8NlWZavPSws7KT7t9vtfbY31XAiy7LU3d3dv8Gcwb6uvvpq/cM//IMWLVqkF198UXa73a/nYNSoUXrttde0fft2vffee/rXf/1XPfTQQ5o0aZKvb0dHh26++WZNnjxZSUlJmjZtmv77v//b+BycOIZZs2YpNzdX0veHih05csR3/0UXXWTc7vXXX9d3332nadOmKS0tTWlpaVqwYIFSU1O1e/fuXm/0vV6vhg0bpqqqKl9bS0uLhg0bpg8//ND4mp7Y1tjYqObm5l6hoqioSEeOHDGGCgD4MfGRBQAEwcSJE7Vu3Tp1dXWpp6dHDz74oFatWnXKbVJTU/Xqq6+qtbVVklRRUeG7LywsrN9v8v/pn/5JNTU1OnTokCRp/fr1Gj58eJ9P6oOxr4SEBN1+++0aNmyYnnrqKUn+PQfPP/+8Fi1apIkTJ2rhwoWaOHGidu/e3atPY2Oj2traNH/+fE2aNEnbt2/3PcaJjn/eJk6cqIqKCrW1tUmSnnzySd1///2nHfvQoUO1atUqff755762r776SmFhYfr5z3+u8PBweb1eWZalSy+9VIMHD/aFhR++daq+vl6S9Mknn+iTTz6R9P05IFdddZVvJekHzc3N+vd//3ffc71x40aNGTOGoABgQGBlAQD6oaOjo8/Xp77wwgv6t3/7N61cuVI333yzvF6vLr/8chUUFJxyX8nJyfrVr36lW2+9VYMHD9aYMWM0ZMgQSdKECRP0hz/8Qb/73e96nXx7Ktdff73uvPNO3XHHHerp6VF0dLT++Mc/+nUoy5nsy2az6dFHH9VNN92k1NRUv56Dm266Se+//76mTp2qIUOGaOTIkX3G+3d/93e64YYblJGRoYiICI0dO1a//OUv1djYqJ///Oe9+qakpGjFihWSpN/85jdqamrSr371K9lsNo0cOdJ336lce+21evDBB/XAAw+otbVVYWFhiomJ0TPPPKOf/vSnioqK0pVXXqnMzEytW7dOpaWlWr58uZ599ll1d3dr3rx5SkxM9H0D1erVq7Vv3z5FR0fr8ccf7/N4SUlJuvvuu5Wfn6+wsDA5HA794Q9/OG2dAPBjsFmnW8cFAJwVH3/8sT744APfd/Y/99xz+uijj7R69eoQV4Zg2L59ux5++GFVV1eHuhQA8BsrCwAQIpdeeqmeeeYZlZeX+z75fvjhh0NdFgAAPqwsAAAAADDiBGcAAAAARoQFAAAAAEbn7DkLPT09am9vl91ul81mC3U5AAAAwDnJsix5PB4NHTq0zzfonbNhob29XZ9++mmoywAAAADOC2PHjtWwYcN6tZ2zYeGHq5mOHTtWERERIaujvr5eCQkJIXt8nJuYN/AH8wb+YN7AH8ybC0tXV5c+/fRT3/vr452zYeGHQ48iIiIUGRkZ0lpC/fg4NzFv4A/mDfzBvIE/mDcXHtOh/ZzgDAAAAMCIsAAAAADAiLAAAAAAwIiwAAAAAMCIsAAAAADAiLAAAAAAwOiMwkJbW5uysrL09ddf92r/y1/+opkzZ/pu79mzR06nU2lpaVqyZIm6u7slSfv371deXp7S09M1Z84ctbe3S5KOHj2q2bNnKyMjQ3l5eWpubg7WuAAAAAAE6LRh4aOPPtKMGTPU0NDQq/3zzz/Xn/70p15tCxcuVGFhoTZv3izLslReXi5JWrZsmXJzc+VyuZSQkKDS0lJJ0urVq5WUlKRNmzbplltu0fLly4M0LAAAAACBOm1YKC8vV1FRkRwOh6+tq6tLhYWFuvfee31t+/btU2dnpyZMmCBJcjqdcrlc8ng82rFjh9LS0nq1S9LWrVuVnZ0tScrKytKbb74pj8cTvNEBAAAA8Ntpr+Bs+rT/iSee0LRp0/Szn/3M13bw4EHFxMT4bsfExKipqUmHDx9WVFSUwsPDe7WfuE14eLiioqJ06NAhjRgxIrBRAQAAAAjYacPCid555x0dOHBAixYt0vbt233tPT09vS4RbVmWbDab7+fxTJeS/mGbQYP6d851fX19v/qfDXV1daEuAecg5g38wbyBP5g38AfzBpIfYaG6ulqfffaZcnJy1NHRoZaWFs2fP18LFy7sdYJyS0uLHA6HoqOj1draKq/Xq7CwMDU3N/sOaXI4HGppaVFsbKy6u7vV3t6u4cOH96uehIQERUZG9ncYQVNXV6fExMSQPT7OTcwb+IN5A38wb+AP5s2Fxe12n/QD+H5/depjjz2mTZs2qaqqSo888ogSEhK0evVqxcXFKTIy0pdCq6qqlJKSIrvdrqSkJNXU1EiSKisrlZKSIklKTU1VZWWlJKmmpkZJSUmy2+1+DRIAAABAcAX1OgvFxcV67LHHlJ6ero6ODuXn50uSioqKVF5erqlTp2rnzp2aP3++JGnevHn68MMPlZmZqeeff16FhYXBLAc4J3V5vCHdHgAA4AdnfBjSli1b+rRdc801uuaaa3y34+PjVVFR0adfXFycysrK+rQPHz5cTz/99JmWAFwQIuxhyr6vyu/tNz6RE8RqAADAhYwrOAMAAAAwIiwAAAAAMCIsAAAAADAiLAAAAAAwIiwAAAAAMCIsAAAAADAiLAAAAAAwIiwAAAAAMCIsAAAAADAiLAAAAAAwIiwAAAAAMCIsAAAAADAiLAAAAAAwIiwAAAAAMCIsAAAAADAiLAAAAAAwIiwAAAAAMCIsAAAAADAiLAAAAAAwIiwAAAAAMCIsAAAAADAiLAAAAAAwIiwAAAAAMCIsAAAAADAiLAAAAAAwIiwAAAAAMCIsAAAAADAiLAAAAAAwIiwAAAAAMCIsAAAAADAiLAAAAAAwIiwAAAAAMCIsAAAAADAiLOC80eXxDoh9AAAAnC/CQ10AECwR9jBl31cV0D42PpETpGoAAADOfawsAAAAADA6o7DQ1tamrKwsff3115KkF198UVlZWcrOztaiRYvU1dUlSdqzZ4+cTqfS0tK0ZMkSdXd3S5L279+vvLw8paena86cOWpvb5ckHT16VLNnz1ZGRoby8vLU3Nx8NsYIAAAAwA+nDQsfffSRZsyYoYaGBknS3r17tXbtWr3wwgvasGGDenp69Pzzz0uSFi5cqMLCQm3evFmWZam8vFyStGzZMuXm5srlcikhIUGlpaWSpNWrVyspKUmbNm3SLbfcouXLl5+lYQIAAADor9OGhfLychUVFcnhcEiSIiIiVFRUpKioKNlsNo0dO1b79+/Xvn371NnZqQkTJkiSnE6nXC6XPB6PduzYobS0tF7tkrR161ZlZ2dLkrKysvTmm2/K4/GclYECAAAA6J/TnuB84qf9cXFxiouLkyQdOnRI69at02OPPaaDBw8qJibG1y8mJkZNTU06fPiwoqKiFB4e3qtdUq9twsPDFRUVpUOHDmnEiBHBGR0AAAAAv/n9bUhNTU2aNWuWpk2bpmuuuUZ1dXWy2Wy++y3Lks1m8/083om3j99m0KD+nXNdX1/f/+KDrK6uLtQlQFJiYmJQ9vNjvZ4ne5xgjIM5ef7itYU/mDfwB/MGkp9h4a9//atmzZqlmTNn6q677pIkxcbG9jpBuaWlRQ6HQ9HR0WptbZXX61VYWJiam5t9hzQ5HA61tLQoNjZW3d3dam9v1/Dhw/tVS0JCgiIjI/0ZRlDU1dUF7U0qBoYf4/U82/OGOXl+4u8N/MG8gT+YNxcWt9t90g/g+/3VqW1tbfr1r3+tefPm+YKC9P3hSZGRkb4UWlVVpZSUFNntdiUlJammpkaSVFlZqZSUFElSamqqKisrJUk1NTVKSkqS3W7vb0kAAAAAzoJ+h4WKigq1tLToueeeU05OjnJycvTkk09KkoqLi/XYY48pPT1dHR0dys/PlyQVFRWpvLxcU6dO1c6dOzV//nxJ0rx58/Thhx8qMzNTzz//vAoLC4M4NAAAAACBOOPDkLZs2SJJuvPOO3XnnXca+8THx6uioqJPe1xcnMrKyvq0Dx8+XE8//fSZlgAAAADgR8QVnAEAAAAYERYAAAAAGBEWAAAAABgRFgAAAAAYERYAAAAAGBEWAAAAABgRFgAAAAAYERYAAAAAGBEWAAAAABgRFgAAAAAYERYAAAAAGBEWAAAAABgRFgAAAAAYERYAAAAAGBEWAAAAABgRFgAAAAAYERYAAAAAGBEWAAAAABgRFgAAAAAYERYAAAAAGBEWAAAAABgRFgAAAAAYERYAAAAAGBEWAAAAABgRFgAAAAAYERYAAAAAGBEWAAAAABgRFgAAAAAYERYAAAAAGBEWAAAAABgRFgAAAAAYERYAAAAAGBEWAAAAABgRFgAAAAAYERYAAAAAGBEWAAAAABidUVhoa2tTVlaWvv76a0lSbW2tsrOzNWXKFJWUlPj67dmzR06nU2lpaVqyZIm6u7slSfv371deXp7S09M1Z84ctbe3S5KOHj2q2bNnKyMjQ3l5eWpubg72+AAAAAD46bRh4aOPPtKMGTPU0NAgSers7NTixYtVWlqqmpoa1dfXa9u2bZKkhQsXqrCwUJs3b5ZlWSovL5ckLVu2TLm5uXK5XEpISFBpaakkafXq1UpKStKmTZt0yy23aPny5WdpmAAAAAD667Rhoby8XEVFRXI4HJKkXbt2afTo0Ro1apTCw8OVnZ0tl8ulffv2qbOzUxMmTJAkOZ1OuVwueTwe7dixQ2lpab3aJWnr1q3Kzs6WJGVlZenNN9+Ux+M5KwMFAAAA0D/hp+tw4qf9Bw8eVExMjO+2w+FQU1NTn/aYmBg1NTXp8OHDioqKUnh4eK/2E/cVHh6uqKgoHTp0SCNGjAh8ZAAAAAACctqwcKKenh7ZbDbfbcuyZLPZTtr+w8/jnXj7+G0GDerfOdf19fX96n821NXVhboESEpMTAzKfn6s1/NkjxOMcTAnz1+8tvAH8wb+YN5A8iMsxMbG9joRubm5WQ6Ho097S0uLHA6HoqOj1draKq/Xq7CwMF9/6ftViZaWFsXGxqq7u1vt7e0aPnx4v+pJSEhQZGRkf4cRNHV1dUF7k4qB4cd4Pc/2vGFOnp/4ewN/MG/gD+bNhcXtdp/0A/h+f3Xq+PHjtXfvXjU2Nsrr9aq6ulopKSmKi4tTZGSkL4VWVVUpJSVFdrtdSUlJqqmpkSRVVlYqJSVFkpSamqrKykpJUk1NjZKSkmS32/0aJAAAAIDg6vfKQmRkpFasWKG5c+fK7XYrNTVV6enpkqTi4mItXbpUbW1tGjdunPLz8yVJRUVFKigo0Jo1azRy5EitWrVKkjRv3jwVFBQoMzNTw4YNU3FxcRCHBgAAACAQZxwWtmzZ4vs9OTlZGzZs6NMnPj5eFRUVfdrj4uJUVlbWp3348OF6+umnz7QEAAAAAD8iruAMAAAAwIiwAAAAAMCIsAAAAADAiLAAAAAAwIiwAAAAAMCIsAAAAADAiLAAAAAAwIiwAAAAAMCIsAAAAADAiLAAAAAAwIiwAAAAAMCIsAAAAADAiLAAAAAAwIiwAAAAAMCIsAAAAADAiLAASVKXxxvS7QEAADDwhIe6AAwMEfYwZd9X5ff2G5/ICWI1AAAAGAhYWQAAAABgRFgAAAAAYERYAAAAAGBEWAAAAABgRFgAAAAAYERYAAAAAGBEWACC6EyvN5GYmHiWKwEAAAgc11kAgijQ61VIXLMCAAAMHKwsAAAAADAiLAAAAAAwIiwAAAAAMCIsAAAAADAiLAAAAAAwIiwAAAAAMCIsAMc50+skAAAAXAi4zgJwnECvk8A1EgAAwPmElQXgPBOM1RFWWAAAgMTKAnDe4SrSAAAgWAJaWaiqqlJmZqYyMzO1cuVKSVJtba2ys7M1ZcoUlZSU+Pru2bNHTqdTaWlpWrJkibq7uyVJ+/fvV15entLT0zVnzhy1t7cHUhIAAACAIPE7LBw7dkzLly9XWVmZqqqqtHPnTm3ZskWLFy9WaWmpampqVF9fr23btkmSFi5cqMLCQm3evFmWZam8vFyStGzZMuXm5srlcikhIUGlpaXBGRkAAACAgPgdFrxer3p6enTs2DF1d3eru7tbUVFRGj16tEaNGqXw8HBlZ2fL5XJp37596uzs1IQJEyRJTqdTLpdLHo9HO3bsUFpaWq92AAAAAKHn9zkLUVFRmjdvnjIyMjRkyBBdffXVOnjwoGJiYnx9HA6Hmpqa+rTHxMSoqalJhw8fVlRUlMLDw3u1AwAAAAg9v8PCJ598ovXr1+uNN97QsGHD9B//8R9qaGiQzWbz9bEsSzabTT09Pcb2H34e78Tbp1NfX+/vEIKmrq4u1CUELDExMeB9hPp5CMYY8P+F+vWEGa8L/MG8gT+YN5ACCAtvv/22kpOTdfHFF0v6/hCitWvXKiwszNenublZDodDsbGxam5u9rW3tLTI4XAoOjpara2t8nq9CgsL8/Xvj4SEBEVGRvo7jIDV1dXxJvX/4Xk4v/B6Djz8vYE/mDfwB/PmwuJ2u0/6Abzf5yzEx8ertrZWHR0dsixLW7Zs0fjx47V37141NjbK6/WqurpaKSkpiouLU2RkpC+hVlVVKSUlRXa7XUlJSaqpqZEkVVZWKiUlxd+SAAAAAASR3ysLEydO1O7du+V0OmW323XFFVdo7ty5uv766zV37ly53W6lpqYqPT1dklRcXKylS5eqra1N48aNU35+viSpqKhIBQUFWrNmjUaOHKlVq1YFZ2Q453R5vIqwh52+IwAAAH4UAV2Ubfbs2Zo9e3avtuTkZG3YsKFP3/j4eFVUVPRpj4uLU1lZWSBl4DwR6MXEuJAYAABAcAV0UTYAAAAA5y/CAgAAAAAjwgIAAAAAI8ICAAAAACPCAgAAAAAjwgIAAAAAI8ICAAAAACPCAgAAAAAjwgIAAAAAI8ICAAAAACPCAgAAAAAjwgIAAAAAI8ICAAAAACPCAgAAAAAjwgIAAAAAI8ICgqLL4w11CQAAAAiy8FAXgPNDhD1M2fdVBbSPjU/kBKkaAAAABAMrCwAAAACMCAsAAAAAjAgLAAAAAIwICwAAAACMCAsAAAAAjAgLAAAAAIwICwAAAACMCAsAAAAAjAgLAAAAAIwICwAAAACMCAsAAAAAjAgLAAAAAIwICwAAAACMCAsAAAAAjAgLAAAAAIwICwAAAACMCAsAAAAAjAgLAAAAAIwICwAAAACMAgoLW7ZskdPpVEZGhh555BFJUm1trbKzszVlyhSVlJT4+u7Zs0dOp1NpaWlasmSJuru7JUn79+9XXl6e0tPTNWfOHLW3twdSEgAAAIAg8TssfPXVVyoqKlJpaak2bNig3bt3a9u2bVq8eLFKS0tVU1Oj+vp6bdu2TZK0cOFCFRYWavPmzbIsS+Xl5ZKkZcuWKTc3Vy6XSwkJCSotLQ3OyAAAAAAExO+w8Nprr2nq1KmKjY2V3W5XSUmJhgwZotGjR2vUqFEKDw9Xdna2XC6X9u3bp87OTk2YMEGS5HQ65XK55PF4tGPHDqWlpfVqBwAAABB64f5u2NjYKLvdrrvvvlsHDhzQDTfcoDFjxigmJsbXx+FwqKmpSQcPHuzVHhMTo6amJh0+fFhRUVEKDw/v1Q4AAAAg9PwOC16vVzt37lRZWZkuuugizZkzR4MHD5bNZvP1sSxLNptNPT09xvYffh7vxNunU19f7+8Qgqauri7UJQQsMTEx1CVggDkf5vX5iNcF/mDewB/MG0gBhIVLLrlEycnJio6OliRNnjxZLpdLYWFhvj7Nzc1yOByKjY1Vc3Ozr72lpUUOh0PR0dFqbW2V1+tVWFiYr39/JCQkKDIy0t9hBKyuro432jgvMa8HHv7ewB/MG/iDeXNhcbvdJ/0A3u9zFm688Ua9/fbbOnr0qLxer9566y2lp6dr7969amxslNfrVXV1tVJSUhQXF6fIyEhfQq2qqlJKSorsdruSkpJUU1MjSaqsrFRKSoq/JQEAAAAIIr9XFsaPH69Zs2YpNzdXHo9H119/vWbMmKFf/OIXmjt3rtxut1JTU5Weni5JKi4u1tKlS9XW1qZx48YpPz9fklRUVKSCggKtWbNGI0eO1KpVq4IzMgB+6/J4FWEPO33Hs7Q9AAAYGPwOC5I0ffp0TZ8+vVdbcnKyNmzY0KdvfHy8Kioq+rTHxcWprKwskDIABFmEPUzZ91X5vf3GJ3KCWA0AAAgVruAMAAAAwIiwAAAAAMCIsAAAAADAiLAAAAAAwIiwAAAAAMCIsAAAAADAiLAAAAAAwIiwAAAAAMCIsAAAAADAiLAAAAAAwIiwAAAAAMCIsAAAAADAiLAAAAAAwIiwAAAAAMCIsAAAAADAiLAAAAAAwIiwAAAAAMCIsAAAAADAiLAAAAAAwIiwAAAAAMCIsAAAAADAiLBwHujyeENdAgAAAM5D4aEuAIGLsIcp+76qgPax8YmcIFUDAACA8wUrCwAAAACMCAsAAAAAjAgLAAAAAIwICwAAAACMCAsAAAAAjAgLAAAAAIwICwAAAACMCAsAAAAAjAgLAAAAAIwICwCCrsvjHRD7AAAAgQkPdQEAzj8R9jBl31cV0D42PpETpGoAAIC/WFkAAAAAYERYAAAAAGAUlLCwcuVKFRQUSJJqa2uVnZ2tKVOmqKSkxNdnz549cjqdSktL05IlS9Td3S1J2r9/v/Ly8pSenq45c+aovb09GCUBAAAACFDAYeHdd9/VK6+8Iknq7OzU4sWLVVpaqpqaGtXX12vbtm2SpIULF6qwsFCbN2+WZVkqLy+XJC1btky5ublyuVxKSEhQaWlpoCUBAAAACIKAwsJ3332nkp8XfNYAAA1eSURBVJIS3X333ZKkXbt2afTo0Ro1apTCw8OVnZ0tl8ulffv2qbOzUxMmTJAkOZ1OuVwueTwe7dixQ2lpab3aAQAAAIReQGGhsLBQCxYs0E9+8hNJ0sGDBxUTE+O73+FwqKmpqU97TEyMmpqadPjwYUVFRSk8PLxXOwAAAIDQ8/urU1966SWNHDlSycnJevnllyVJPT09stlsvj6WZclms520/Yefxzvx9unU19f7O4SgqaurC+njJyYmhvTxgbMl1P+2BiKeE/iDeQN/MG8gBRAWampq1NzcrJycHB05ckQdHR3at2+fwsLCfH2am5vlcDgUGxur5uZmX3tLS4scDoeio6PV2toqr9ersLAwX//+SEhIUGRkpL/DCFhdXR1v1oGzhH9bvfH3Bv5g3sAfzJsLi9vtPukH8H4fhvTcc8+purpaVVVVuvfeezVp0iQ9++yz2rt3rxobG+X1elVdXa2UlBTFxcUpMjLSl1CrqqqUkpIiu92upKQk1dTUSJIqKyuVkpLib0kAAAAAgiioV3COjIzUihUrNHfuXLndbqWmpio9PV2SVFxcrKVLl6qtrU3jxo1Tfn6+JKmoqEgFBQVas2aNRo4cqVWrVgWzJAAAAAB+CkpYcDqdcjqdkqTk5GRt2LChT5/4+HhVVFT0aY+Li1NZWVkwygAAAAAQRFzBGQAAAIARYQEAAACAEWEBAAAAgBFhAQAAAIARYQEAAACAEWFhAOjyeENdAgAAANBHUK+zAP9E2MOUfV+V39tvfCIniNUAAAAA32NlAQAAAIARYQEAAACAEWEBAAAAgBFhIUDxl48LdQkAAADAWcEJzgEaetHggE5OljhBGQAAAAMTKwsAAAAAjAgLAAAAAIwICwAAAACMCAsAAAAAjAgLAAAAAIwICwAAAACMCAsAAAAAjAgLAAAAAIwICwAAAACMCAsAAAAAjAgLAAAAAIwICwAAAACMCAsAAAAAjAgLAAAAAIwICwAAAACMCAsAAAAAjAgLAAAAAIwICwAAAACMCAsAAAAAjAgLAAAAAIwICwAAAACMCAsAAAAAjAgLAAAAAIwICwAAAACMAgoLTz31lDIzM5WZmanHH39cklRbW6vs7GxNmTJFJSUlvr579uyR0+lUWlqalixZou7ubknS/v37lZeXp/T0dM2ZM0ft7e2BlAQAAAAgSPwOC7W1tXr77bf1yiuvqLKyUv/zP/+j6upqLV68WKWlpaqpqVF9fb22bdsmSVq4cKEKCwu1efNmWZal8vJySdKyZcuUm5srl8ulhIQElZaWBmdkAAAAAALid1iIiYlRQUGBIiIiZLfbddlll6mhoUGjR4/WqFGjFB4eruzsbLlcLu3bt0+dnZ2aMGGCJMnpdMrlcsnj8WjHjh1KS0vr1Q4AAAAg9PwOC2PGjPG9+W9oaNCmTZtks9kUExPj6+NwONTU1KSDBw/2ao+JiVFTU5MOHz6sqKgohYeH92oHAAAAEHrhge7gs88+029/+1vdf//9CgsLU0NDg+8+y7Jks9nU09Mjm83Wp/2Hn8c78fbp1NfXB1R/oBITE0P6+MD5rK6uLtQlDDg8J/AH8wb+YN5ACjAs1NXV6d5779XixYuVmZmp999/X83Nzb77m5ub5XA4FBsb26u9paVFDodD0dHRam1tldfrVVhYmK9/fyQkJCgyMjKQYQAYoAjjvdXV1fGcoN+YN/AH8+bC4na7T/oBvN+HIR04cED33HOPiouLlZmZKUkaP3689u7dq8bGRnm9XlVXVyslJUVxcXGKjIz0JdSqqiqlpKTIbrcrKSlJNTU1kqTKykqlpKT4WxIAAACAIPJ7ZWHt2rVyu91asWKFr+22227TihUrNHfuXLndbqWmpio9PV2SVFxcrKVLl6qtrU3jxo1Tfn6+JKmoqEgFBQVas2aNRo4cqVWrVgU4JAAAAADB4HdYWLp0qZYuXWq8b8OGDX3a4uPjVVFR0ac9Li5OZWVl/pYBAAAA4CzhCs4AAAAAjAgLAAAAAIwICwAAAACMCAsAAAAAjAgLAAakLo83pNsDAIAgXMEZAM6GCHuYsu+r8nv7jU/kBLEaAAAuTKwsAAAAADAiLADASXAoFADgQsdhSABwEhwKBQC40LGyAAAAAMCIsAAAAADAiLAAAAAAwIiwAAAAAMCIsAAAAADAiLAA4LzE15YCABA4vjoVwHkp0K89lfjqUwAAWFkAAAAAYERYAIDzWDAOx+KQLgC4cHEYEgCcxzgcCwAQCFYWAAAAABgRFgAAAAAYERYAAAAAGBEWAOAs4eRiAMC5jhOcAeAsCcbJxetXZPl+T0xMDLQkAAD6hbAAAANYoIGDbzICAASCw5AAAKcU6KFQHEoFAOcuVhYAAKc0EFY3ujxeRdjDQrY9AFyoCAsAgAFvIAQWALgQcRgSAOCs4jAkADh3sbIAADirgvGtUKwMAEBosLIAAAAAwIiwAAAAAMCIsAAAAADAiLAAADjvca0IAPAPJzgDAM57fPUqAPhnQKwsbNy4UVOnTtWUKVO0bt26UJcDAEAvwVhZYHUCwLko5CsLTU1NKikp0csvv6yIiAjddtttuuaaa/TLX/4y1KUBACApOF//un5FVkDbcxVqAKEQ8rBQW1ura6+9VsOHD5ckpaWlyeVy6Xe/+12IKwMAIHg4FArAuSjkYeHgwYOKiYnx3XY4HNq1a9dpt7MsS5LU1dV11mo7U8OHBvZJj9vtDmgfgW5PDdRADdRADWd3+2Dso629Q/bw/m+fkJAgt9stSfJ0e/3aBy5MP8wbnP9+eD/9w/vr49ksU+uPaM2aNXK73Zo/f74kqby8XPX19XrooYdOuV1ra6s+/fTTH6NEAAAA4Lw3duxYDRs2rFdbyFcWYmNjtXPnTt/t5uZmORyO0243dOhQjR07Vna7XTab7WyWCAAAAJy3LMuSx+PR0KFD+9wX8rBw3XXX6fe//70OHTqkIUOG6NVXX9XDDz982u0GDRrUJ/kAAAAA6L/Bgwcb20MeFkaMGKEFCxYoPz9fHo9H06dP15VXXhnqsgAAAIALXsjPWQAAAAAwMA2Ii7IBAAAAGHgICwAAAACMCAsAAAAAjAgLAAAAAIwICwAAAACMCAsB2Lhxo6ZOnaopU6Zo3bp1oS4HA9RTTz2lzMxMZWZm6vHHH5ck1dbWKjs7W1OmTFFJSUmIK8RAtnLlShUUFEhi3uD0tmzZIqfTqYyMDD3yyCOSmDc4vaqqKt//UytXrpTEvMFxLPjlm2++sW688Ubr8OHDVnt7u5WdnW199tlnoS4LA8w777xj3XrrrZbb7ba6urqs/Px8a+PGjVZqaqr15ZdfWh6Px7rrrrusrVu3hrpUDEC1tbXWNddcYz3wwAPWsWPHmDc4pS+//NKaOHGideDAAaurq8uaMWOGtXXrVuYNTqmjo8O6+uqrrW+//dbyeDzW9OnTrddff515Ax9WFvxUW1ura6+9VsOHD9dFF12ktLQ0uVyuUJeFASYmJkYFBQWKiIiQ3W7XZZddpoaGBo0ePVqjRo1SeHi4srOzmTvo47vvvlNJSYnuvvtuSdKuXbuYNzil1157TVOnTlVsbKzsdrtKSko0ZMgQ5g1Oyev1qqenR8eOHVN3d7e6u7sVFRXFvIFPyK/gfK46ePCgYmJifLcdDod27doVwoowEI0ZM8b3e0NDgzZt2qTbb7+9z9xpamoKRXkYwAoLC7VgwQIdOHBAkvlvDvMGx2tsbJTdbtfdd9+tAwcO6IYbbtCYMWOYNzilqKgozZs3TxkZGRoyZIiuvvpq/t6gF1YW/NTT0yObzea7bVlWr9vA8T777DPddddduv/++zVq1CjmDk7ppZde0siRI5WcnOxr428OTsfr9erdd9/Vo48+qhdffFG7du3SV199xbzBKX3yySdav3693njjDb311lsaNGiQGhoamDfwYWXBT7Gxsdq5c6fvdnNzsxwORwgrwkBVV1ene++9V4sXL1ZmZqbef/99NTc3++5n7uBENTU1am5uVk5Ojo4cOaKOjg7t27dPYWFhvj7MG5zokksuUXJysqKjoyVJkydPlsvlYt7glN5++20lJyfr4osvliQ5nU6tXbuWeQMfVhb8dN111+ndd9/VoUOHdOzYMb366qtKSUkJdVkYYA4cOKB77rlHxcXFyszMlCSNHz9ee/fuVWNjo7xer6qrq5k76OW5555TdXW1qqqqdO+992rSpEl69tlnmTc4pRtvvFFvv/22jh49Kq/Xq7feekvp6enMG5xSfHy8amtr1dHRIcuytGXLFv6fQi+sLPhpxIgRWrBggfLz8+XxeDR9+nRdeeWVoS4LA8zatWvldru1YsUKX9ttt92mFStWaO7cuXK73UpNTVV6enoIq8S5IDIyknmDUxo/frxmzZql3NxceTweXX/99ZoxY4Z+8YtfMG9wUhMnTtTu3bvldDplt9t1xRVXaO7cubr++uuZN5Ak2SzLskJdBAAAAICBh8OQAAAAABgRFgAAAAAYERYAAAAAGBEWAAAAABgRFgAAAAAYERYAAAAAGBEWAAAAABgRFgAAAAAY/R+WobqbbKEWEQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 936x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "4. Applied automispelling corrections. Shape is still (106648,). \n",
      " \n",
      " Peek: \n",
      " 0    [the, newest, update, i️, made, sure, to, down...\n",
      "1                                                   []\n",
      "2    [tried, resetting, my, settings, restarting, m...\n",
      "3                  [this, is, what, it, looks, like, ]\n",
      "4    [i️, have, an, iphone, 7, plus, and, yes, i️, do]\n",
      "Name: inbound_text, dtype: object\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "1b06c7cba6864788a6788e8674d3944a",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=106648.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "5. Removed the stop words. Shape is still (106648,). \n",
      " \n",
      " Peek: \n",
      " 0    [newest, update, i️, made, sure, download, yes...\n",
      "1                                                   []\n",
      "2      [tried, resetting, settings, restarting, phone]\n",
      "3                                      [looks, like, ]\n",
      "4                       [i️, iphone, 7, plus, yes, i️]\n",
      "Name: inbound_text, dtype: object\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 936x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "5859af7ddb9d4da1802f3bf5baf2bcde",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=106648.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "6. Expanded contractions into extra tokens. Shape is still (106648,). \n",
      " \n",
      " Peek: \n",
      " 0    [newest, update, i️, made, sure, download, yes...\n",
      "1                                                   []\n",
      "2      [tried, resetting, settings, restarting, phone]\n",
      "3                                      [looks, like, ]\n",
      "4                       [i️, iphone, 7, plus, yes, i️]\n",
      "Name: inbound_text, dtype: object\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 936x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/anaconda3/envs/chatbot/lib/python3.7/site-packages/ipykernel_launcher.py:54: TqdmDeprecationWarning: This function will be removed in tqdm==5.0.0\n",
      "Please use `tqdm.notebook.tqdm` instead of `tqdm.tqdm_notebook`\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "3c9bb949b30b421ba602e6ce4bdfba91",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=1.0, bar_style='info', max=1.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "7. Remove all non-english Tweets. Shape is now (93391,). Clearly less than before! \n",
      " \n",
      " Peek: \n",
      " 0    [newest, update, i️, made, sure, download, yes...\n",
      "2      [tried, resetting, settings, restarting, phone]\n",
      "4                       [i️, iphone, 7, plus, yes, i️]\n",
      "5                     [i️, need, answers, annoying, 🙃]\n",
      "6    [hey, anyone, else, upgraded, ios, 11.1, issue...\n",
      "Name: inbound_text, dtype: object\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "6ea8f46b7df7482ba514eb66821e585e",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=93391.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "8. Lemmatized the tokens. Shape is still (93391,). \n",
      " \n",
      " Peek: \n",
      " 0    [new, update, i️, make, sure, download, yester...\n",
      "2                [try, reset, setting, restart, phone]\n",
      "4                       [i️, iphone, 7, plus, yes, i️]\n",
      "5                         [i️, need, answer, annoy, 🙃]\n",
      "6    [hey, anyone, else, upgraded, io, 11.1, issue,...\n",
      "Name: inbound_text, dtype: object\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 936x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "0268e1e9433b479c9d086f231776dab7",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=93391.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "da3d1b13521e46268bba92666805cefc",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=93391.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "9. Removed emojis and numbers. Shape is still (93391,). \n",
      " \n",
      " Peek: \n",
      " 0    [new, update, i️, make, sure, download, yester...\n",
      "2                [try, reset, setting, restart, phone]\n",
      "4                          [i️, iphone, plus, yes, i️]\n",
      "5                            [i️, need, answer, annoy]\n",
      "6    [hey, anyone, else, upgraded, io, issue, capit...\n",
      "Name: inbound_text, dtype: object\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 936x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "05abfd05400e4dafbc86d4303d23776a",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=93391.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "10. Limited each tweet to a max of 50 tokens and a min of 5 tokens. Shape is now (76066,). \n",
      " \n",
      " Peek: \n",
      " 0     [new, update, i️, make, sure, download, yester...\n",
      "6     [hey, anyone, else, upgraded, io, issue, capit...\n",
      "12    [hello, internet, someone, explain, symbol, ke...\n",
      "13    [get, screenshot, say, iphonex, reserve, email...\n",
      "15    [thank, update, phone, even, slow, barely, wor...\n",
      "Name: inbound_text, dtype: object\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 936x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "processed_inbound = my_tokenizer(text['inbound'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This is great, we now have the preprocessed data stored into the variable ```processed_inbound```. Now let's convert this processed data to pkl so that I don't have to run this over and over again to use at the next notebook."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0     [new, update, i️, make, sure, download, yester...\n",
       "6     [hey, anyone, else, upgraded, io, issue, capit...\n",
       "12    [hello, internet, someone, explain, symbol, ke...\n",
       "13    [get, screenshot, say, iphonex, reserve, email...\n",
       "15    [thank, update, phone, even, slow, barely, wor...\n",
       "Name: inbound_text, dtype: object"
      ]
     },
     "execution_count": 92,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "processed_inbound.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "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>Processed Inbound</th>\n",
       "      <th>Real Inbound</th>\n",
       "      <th>Real Outbound</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>[new, update, i️, make, sure, download, yester...</td>\n",
       "      <td>@AppleSupport The newest update. I️ made sure ...</td>\n",
       "      <td>@115854 Lets take a closer look into this issu...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>[hey, anyone, else, upgraded, io, issue, capit...</td>\n",
       "      <td>Hey @AppleSupport and anyone else who upgraded...</td>\n",
       "      <td>@115856 Hey, let's work together to figure out...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>[hello, internet, someone, explain, symbol, ke...</td>\n",
       "      <td>Hello, internet. Can someone explain why this ...</td>\n",
       "      <td>@115861 You're in the right place; we'll do al...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>[get, screenshot, say, iphonex, reserve, email...</td>\n",
       "      <td>@AppleSupport I’ve got a screenshot saying my ...</td>\n",
       "      <td>@115863 Go ahead and send us a DM please. Let ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>[thank, update, phone, even, slow, barely, wor...</td>\n",
       "      <td>Thank you @AppleSupport I updated my phone and...</td>\n",
       "      <td>@115864 We'd like to help, but we'll need more...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                    Processed Inbound  \\\n",
       "0   [new, update, i️, make, sure, download, yester...   \n",
       "6   [hey, anyone, else, upgraded, io, issue, capit...   \n",
       "12  [hello, internet, someone, explain, symbol, ke...   \n",
       "13  [get, screenshot, say, iphonex, reserve, email...   \n",
       "15  [thank, update, phone, even, slow, barely, wor...   \n",
       "\n",
       "                                         Real Inbound  \\\n",
       "0   @AppleSupport The newest update. I️ made sure ...   \n",
       "6   Hey @AppleSupport and anyone else who upgraded...   \n",
       "12  Hello, internet. Can someone explain why this ...   \n",
       "13  @AppleSupport I’ve got a screenshot saying my ...   \n",
       "15  Thank you @AppleSupport I updated my phone and...   \n",
       "\n",
       "                                        Real Outbound  \n",
       "0   @115854 Lets take a closer look into this issu...  \n",
       "6   @115856 Hey, let's work together to figure out...  \n",
       "12  @115861 You're in the right place; we'll do al...  \n",
       "13  @115863 Go ahead and send us a DM please. Let ...  \n",
       "15  @115864 We'd like to help, but we'll need more...  "
      ]
     },
     "execution_count": 98,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Index matching\n",
    "real_inbound = text['inbound_text'].loc[processed_inbound.index]\n",
    "real_outbound = text['outbound_text'].loc[processed_inbound.index]\n",
    "processed_inbound\n",
    "\n",
    "# Storing in DataFrame\n",
    "processed = pd.DataFrame({'Processed Inbound': processed_inbound, 'Real Inbound': real_inbound,\n",
    "             'Real Outbound': real_outbound})\n",
    "\n",
    "# Preview\n",
    "processed.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Saving processed inbound data into a serialized saved file\n",
    "processed_inbound.to_pickle('objects/processed_inbound.pkl')\n",
    "\n",
    "# Saving all processed data\n",
    "processed.to_pickle('objects/processed.pkl')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0     [new, update, i️, make, sure, download, yester...\n",
       "4     [hey, anyone, else, upgraded, io, issue, capit...\n",
       "9     [hello, internet, someone, explain, symbol, ke...\n",
       "10    [get, screenshot, say, iphonex, reserve, email...\n",
       "12    [thank, update, phone, even, slow, barely, wor...\n",
       "dtype: object"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Loading it back in\n",
    "processed_inbound = pd.read_pickle('objects/processed_inbound.pkl')\n",
    "processed_inbound.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Unit testing for preprocessing pipeline, with brief explanations\n",
    "\n",
    "Feel free to skip through this part if you would like. This is where I individually tested each preprocessing step above and commented on some considerations I made at each step. At the end of this section, I also briefly show some other things I considered when I preprocessed my data."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 1. Removing numbers"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "I also want to make sure mine takes out 11.1 because that's just noise."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "True"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "'1'.isdigit()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0         [new, update, i️, make, sure, download, yester...\n",
       "2             [try, reset, setting, .., restart, phone, ..]\n",
       "4                            [i️, iphone, 7, plus, yes, i️]\n",
       "6         [hey, anyone, else, upgraded, io, 11.1, issue,...\n",
       "10                   [hello, line, close, tonight, #, help]\n",
       "                                ...                        \n",
       "106643    [anyone, issue, #, osx, #, highsierra, #, slac...\n",
       "106644    [hey, able, duplicate, file, #, page, search, ...\n",
       "106645    [yo, weird, glitch, w, capital, i️, attempt, m...\n",
       "106646    [fuck, phone, keep, hang, call, show, call, fa...\n",
       "106647    [anyone, #, iphone, issue, phone, freeze, rand...\n",
       "Name: inbound_text, Length: 79760, dtype: object"
      ]
     },
     "execution_count": 62,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "processed_inbound"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['i', 'dont', 'want', 'numbers']"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test = ['1', 'i', 'dont', 'want', 'numbers', '12', '44','11.1']\n",
    "\n",
    "no_ints = [x for x in test if not any(c.isdigit() for c in x)]\n",
    "\n",
    "no_ints"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Nested list comprehensions to apply it to the entire series\n",
    "# [[x for x in i if not any(c.isdigit() for c in x)] for i in processed_inbound]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 2. Testing out the TweetTokenizer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "test = \" \".join(['@KingJames','I','am', 'the', '#GOAT'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['I', 'am', 'the', '#GOAT']"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tknzr = TweetTokenizer(strip_handles = True, reduce_len = True)\n",
    "# Using progress apply to show the progress bar\n",
    "tknzr.tokenize(test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "So now I know that this tokenizer does not take out #, so I will add it on "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3. Removing stuff from a list "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "metadata": {},
   "outputs": [],
   "source": [
    "test = ['', 'bye', 'empty', 'string']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "metadata": {},
   "outputs": [],
   "source": [
    "def remove_from_list(x, stuff_to_remove):\n",
    "    ''' Making a function to remove a list of items from a list'''\n",
    "    for item in stuff_to_remove:\n",
    "        # Making sure to iterate through the entire token\n",
    "        for i,token in enumerate(x):\n",
    "            if item == token:\n",
    "                del x[i]\n",
    "    return x"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['bye', 'empty', 'string']"
      ]
     },
     "execution_count": 95,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "remove_from_list(test, puncts)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3.1 Removing links\n",
    "Going to be doing this with a regular expression."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [],
   "source": [
    "test = processed_X_train[0:10]\n",
    "test;"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "def remove_links(doc):\n",
    "    return [re.sub(r'^https?:\\/\\/.*[\\r\\n]*', '', token, flags=re.MULTILINE) for token in doc]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [],
   "source": [
    "test.apply(remove_links);"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 4. Testing out the Spell Checker"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['can', 'you', 'two', 'to', 'spell', 'check', 'and', 'fun', 'this', 'ios']"
      ]
     },
     "execution_count": 86,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test = ['Cna','you', 'twy', 'to', 'spel', 'check','and','fux', 'tihs','ios']\n",
    "\n",
    "spell = SpellChecker()\n",
    "def correct_spellings(x):\n",
    "    ''' Takes as input a list and outputs a list of the corrected spelling'''\n",
    "    corrected_text = []\n",
    "    for word in x:\n",
    "        if word in x:\n",
    "            corrected_text.append(spell.correction(word))\n",
    "        else:\n",
    "            corrected_text.append(word)\n",
    "    return corrected_text\n",
    "        \n",
    "text = \"speling correctin\"\n",
    "correct_spellings(test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "So here we see that the function generally does a good job, however it miscorrected twy as try and fux as fix because it doesn't understand context and uses Levianathon distance as it's measure of incorrectly spelt words, but that's okay and generally I think it'll do a good job as a preprocessing step. But it is too slow."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 6. Testing contraction and colloquial langauge expansion step"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 178,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Amount of words replaced: 1\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "0        is\n",
       "1       not\n",
       "2    nobody\n",
       "3     loves\n",
       "4       you\n",
       "dtype: object"
      ]
     },
     "execution_count": 178,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def replace_from_dict(x,dic):\n",
    "    ''' Making a function to replace all the items in a list based on a dictionary. I made sure to use the insert\n",
    "        method to insert the longer-gram replacement as distinct items in the list at that specific index'''\n",
    "    x = x.str.lower() \n",
    "    x = list(x)\n",
    "    replaced_counter = 0\n",
    "    for item in dic.items():\n",
    "        for i, e in enumerate(x):\n",
    "            if e == item[0]:\n",
    "                replaced_counter+=1\n",
    "                # Inserting the expanded tokens in a way that preserves the order\n",
    "                del x[i]\n",
    "                for ix, token in enumerate(item[1].split()):\n",
    "                    x.insert(i+ix,token)\n",
    "    print(f\"Amount of words replaced: {replaced_counter}\")\n",
    "    return pd.Series(x)\n",
    "\n",
    "trial = pd.Series([\"Ain't\",\"nobody\",\"is\",\"as\", \"cute\", \"as\", \"you\"])\n",
    "replace_from_dict(trial,contraction_dict)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 7. Testing out removing non-english Tweets\n",
    "I will be using an extension in the spaCy-universe called spaCy-CLD, which uses Naive Bayes and n-grams as a language classifer. That's pretty cool. Good thing I only have to load it in as a package here! I think it's also more efficient than a language detection module I tried earlier, which took me 2+ hours to apply to my data! Gotta love the work they did for optimization.\n",
    "\n",
    "Extension attributes in spaCy allow you store additional information that are relevant to the application. It's quite a powerful tool. \n",
    "\n",
    "Todo: See what the langdetector was trained on."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [],
   "source": [
    "# I will assume that I already initialized the spaCy objects, as shown in the code cell up top"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [],
   "source": [
    "def only_english(x):\n",
    "    ''' Making a function that only accepts English by appending True if it is English and False otherwise \n",
    "        into a mask. Returns a mask'''    \n",
    "    mask = []\n",
    "    x = x.apply(\" \".join)\n",
    "    try:\n",
    "        for i,doc in tqdm(enumerate(nlp_cld.pipe(x, batch_size=512))):\n",
    "            if 'en' not in doc._.languages or len(doc._.languages) != 1:\n",
    "                mask.append(False)\n",
    "            else:\n",
    "                mask.append(True)\n",
    "    except Exception as e:\n",
    "        print(f\"Exception:{e}\")\n",
    "    return mask"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "@AppleSupport The newest update. I️ made sure to download it yesterday.\n",
      "['en']\n",
      "{'en': 0.98}\n"
     ]
    }
   ],
   "source": [
    "# Demonstrating and experimenting what a doc object is\n",
    "doc = nlp_cld(X_train[0])\n",
    "print(doc)\n",
    "print(doc._.languages)  \n",
    "print(doc._.language_scores)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 8. Testing out the WordNetLemmatizer "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This following lemmatizer tries to break down the word down to it's lemma or dictionary form."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['the', 'stripe', 'bat', 'be', 'hang', 'on', '-PRON-', 'foot', 'for', 'good']"
      ]
     },
     "execution_count": 59,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Initialize spacy 'en' model, keeping only tagger component needed for lemmatization\n",
    "nlp = spacy.load('en', disable=['parser', 'ner'])\n",
    "\n",
    "sentence = \"James is good, but Eko beats him at coding Python so is he really?\"\n",
    "\n",
    "# Parse the sentence using the loaded 'en' model object `nlp`\n",
    "doc = nlp(sentence)\n",
    "\n",
    "# Extract the lemma for each token and join\n",
    "[token.lemma_ for token in doc]\n",
    "#> 'the strip bat be hang on -PRON- foot for good'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Lemmatize with POS Tag\n",
    "\n",
    "def get_wordnet_pos(word):\n",
    "    \"\"\"Map POS tag to first character lemmatize() accepts\"\"\"\n",
    "    tag = nltk.pos_tag([word])[0][1][0].upper()\n",
    "    tag_dict = {\"J\": wordnet.ADJ,\n",
    "                \"N\": wordnet.NOUN,\n",
    "                \"V\": wordnet.VERB,\n",
    "                \"R\": wordnet.ADV}\n",
    "    return tag_dict.get(tag, wordnet.NOUN)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'n'"
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# This function returns the POS\n",
    "get_wordnet_pos('cool')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "foot\n",
      "['James', 'be', 'good', 'at', 'cod', 'python', ',', 'but', 'Eko', 'be', 'even', 'well', 'at', 'cod', 'it']\n"
     ]
    }
   ],
   "source": [
    "# 1. Init Lemmatizer\n",
    "lemmatizer = WordNetLemmatizer()\n",
    "\n",
    "# 2. Lemmatize Single Word with the appropriate POS tag\n",
    "word = 'feet'\n",
    "print(lemmatizer.lemmatize(word, get_wordnet_pos(word)))\n",
    "\n",
    "# 3. Lemmatize a Sentence with the appropriate POS tag\n",
    "sentence = \"James is good at coding python, but Eko is even better at coding\"\n",
    "print([lemmatizer.lemmatize(w, get_wordnet_pos(w)) for w in nltk.word_tokenize(sentence)])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "lemmatizer = nltk.stem.WordNetLemmatizer()\n",
    "w_tokenizer = nltk.tokenize.WhitespaceTokenizer()\n",
    "\n",
    "test = ['This','is','a','row']\n",
    "\n",
    "def lemmatize_text(x):\n",
    "    ''' This lemmatizer f`unction should work on a single list of tokenized data'''\n",
    "    # Turning list into a string\n",
    "    x = \" \".join(x)\n",
    "    # Returning a list again\n",
    "    return [lemmatizer.lemmatize(w, get_wordnet_pos(w)) for w in nltk.word_tokenize(x)]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Failed to catch the past tense of a verb."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "2         [try, reset, setting, .., restart, phone, ..]\n",
       "6     [hey, anyone, else, upgraded, io, 11.1, issue,...\n",
       "10               [hello, line, close, tonight, #, help]\n",
       "12    [hello, internet, someone, explain, symbol, ke...\n",
       "13    [get, screenshot, say, #, iphonex, reserve, 3r...\n",
       "15    [thank, update, phone, even, slow, barely, wor...\n",
       "19    [need, software, update, urgently, battery, la...\n",
       "21         [hi, go, youtube, lose, do, thanks, support]\n",
       "25    [hey, last, time, download, update, freak, pho...\n",
       "Name: inbound_text, dtype: object"
      ]
     },
     "execution_count": 48,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "trial = processed_inbound.iloc[1:10]\n",
    "trial"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "0d4b85574124439c86dca132336bbcaf",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=9.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "26943    [work, apple, support, get, result, paid, deve...\n",
       "18403    [many, stupid, bug, new, update, ready, throw,...\n",
       "19883    [take, 7, plus, charger, get, ready, work, tim...\n",
       "87560    [log, itunes, account, lap-fucking-top, god, g...\n",
       "23736            [oh, sorry, meant, attach, screenshot, 🙄]\n",
       "72623    [ever, since, update, iphone, bluetooth, stay,...\n",
       "11613                [fuck, phone, say, sim, need, answer]\n",
       "6323     [think, there, be, bug, io, 11, someone, sends...\n",
       "66978    [nothing, work, last, update., can, not, open,...\n",
       "Name: inbound_text, dtype: object"
      ]
     },
     "execution_count": 100,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "trial.progress_apply(lemmatize_text)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "iOS is still lemmatized wrong, but that is alright because all the instances of iOS will still be together."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 9. Removing Emojis"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "6    [😤, 🙍, 🎲, 🙃, get, rid, of, 😐, emojis]\n",
       "9                        [This, No, Emoji]\n",
       "dtype: object"
      ]
     },
     "execution_count": 107,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test = pd.Series([['😤', '🙍','🎲', '🙃', 'get', 'rid', 'of', '😐','emojis'],['This', 'No','Emoji']], index = [6,9])\n",
    "test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Redefining\n",
    "def remove_from_list(x, stuff_to_remove):\n",
    "    ''' Making a function to remove a list of items from a list'''\n",
    "    for item in stuff_to_remove:\n",
    "        # Making sure to iterate through the entire token\n",
    "        for i,token in enumerate(x):\n",
    "            if item == token:\n",
    "                del x[i]\n",
    "    return x"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "metadata": {},
   "outputs": [],
   "source": [
    "import emoji\n",
    "\n",
    "def extract_emojis(s):\n",
    "    return [c for c in s if c in emoji.UNICODE_EMOJI]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 116,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "6    [get, rid, of, emojis]\n",
       "9         [This, No, Emoji]\n",
       "dtype: object"
      ]
     },
     "execution_count": 116,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "unique_emojis = [i[0] for i in emoji.UNICODE_EMOJI]\n",
    "test.apply(remove_from_list, stuff_to_remove = unique_emojis)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 10. Limiting the length of each tweet\n",
    "Based on the distribution of the tokens we have been seeing at each step above, I think a good cutoff point would be at 50 tokens. I am going to try something like 75 or 100 because you generally don't want to drop too much data. If you don't drop the data, when you use Keras tokenizer and pad sequences, if it's longer than 100 words it will stop there.\n",
    "\n",
    "Similarly, I want to filter out documents with a number of tokens that are less than a particular threshold."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    [test, test, test, test, test, test, test, tes...\n",
       "1                                       [Another, one]\n",
       "dtype: object"
      ]
     },
     "execution_count": 63,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Creating toy list of length 100\n",
    "test = ['test'] * 100\n",
    "test = pd.Series([test, ['Another','one']])\n",
    "test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [],
   "source": [
    "def limit_length(x, max_tokens, min_tokens):\n",
    "    ''' Inputs a list and drops it out of the document if \n",
    "    it has more than the max and less than the min'''\n",
    "    output = x\n",
    "    if len(x) > max_tokens:\n",
    "        output = np.nan\n",
    "    if len(x) < thres:\n",
    "        output = np.nan\n",
    "    return output"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Alternative Steps\n",
    "Ones that I did not include in my Pipeline for a specific reason, but I still employed or considered."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 2. Cleaning the numbers with regular expressions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [],
   "source": [
    "test = pd.Series([['1','iPhone','is','2','little'],['help','1','please']])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [],
   "source": [
    "def clean_numbers(x):\n",
    "    for i,j in enumerate(x):\n",
    "        if bool(re.search(r'\\d', j)):\n",
    "            del x[i]\n",
    "    return x"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    [iPhone, is, little]\n",
       "1          [help, please]\n",
       "dtype: object"
      ]
     },
     "execution_count": 82,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test = test.apply(clean_numbers)\n",
    "test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0         1\n",
       "1    iPhone\n",
       "2        is\n",
       "3         2\n",
       "4    little\n",
       "dtype: object"
      ]
     },
     "execution_count": 65,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3. N Grams\n",
    "Pairs are always better usually."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Other Datasets I Explored:\n",
    "\n",
    "These are just the two other datasets that I explored. I did not use the Ubuntu one because I did not like the language that the customer's spoke here and I don't think it is not ideal to merge it with by my Twitter Apple inbound data.\n",
    "\n",
    "As for the RSICS data, it's interesting because it also had some data that encodes some relational information of the customer (I ultimately want my chatbot to have a relational aspect to it as well, so it is not a bland chatbot), but I am not interested in travel and plane customer data so I did not use this."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Ubuntu"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "dialogueText.csv     dialogueText_301.csv\r\n",
      "dialogueText_196.csv toc.csv\r\n"
     ]
    }
   ],
   "source": [
    "!ls data/ubuntu"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Reading in Ubuntu data\n",
    "ubuntu_df = pd.read_csv('data/ubuntu/dialogueText.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Visualizing the tables, applying styling so that the table is more readable\n",
    "show = lambda x: x.tail(50).style.set_properties(**{'background-color': 'black',                                                   \n",
    "                                    'color': 'lawngreen',                       \n",
    "                                    'border-color': 'white'})\\\n",
    ".applymap(lambda x: f\"color: {'lawngreen' if isinstance(x,str) else 'red'}\")\\\n",
    ".background_gradient(cmap='Blues')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style  type=\"text/css\" >\n",
       "    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow0_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "            background-color:  #f7fbff;\n",
       "            color:  #000000;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow0_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow0_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow0_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow0_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow0_col5 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow1_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "            background-color:  #f7fbff;\n",
       "            color:  #000000;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow1_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow1_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow1_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow1_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow1_col5 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow2_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "            background-color:  #f7fbff;\n",
       "            color:  #000000;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow2_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow2_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow2_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow2_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow2_col5 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow3_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "            background-color:  #f7fbff;\n",
       "            color:  #000000;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow3_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow3_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow3_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow3_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow3_col5 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow4_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "            background-color:  #f7fbff;\n",
       "            color:  #000000;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow4_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow4_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow4_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow4_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow4_col5 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow5_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "            background-color:  #f7fbff;\n",
       "            color:  #000000;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow5_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow5_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow5_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow5_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow5_col5 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow6_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "            background-color:  #f7fbff;\n",
       "            color:  #000000;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow6_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow6_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow6_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow6_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow6_col5 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow7_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "            background-color:  #f7fbff;\n",
       "            color:  #000000;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow7_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow7_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow7_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow7_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow7_col5 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow8_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "            background-color:  #f7fbff;\n",
       "            color:  #000000;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow8_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow8_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow8_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow8_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow8_col5 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow9_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "            background-color:  #f7fbff;\n",
       "            color:  #000000;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow9_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow9_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow9_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow9_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow9_col5 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow10_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "            background-color:  #f7fbff;\n",
       "            color:  #000000;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow10_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow10_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow10_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow10_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow10_col5 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow11_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "            background-color:  #f7fbff;\n",
       "            color:  #000000;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow11_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow11_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow11_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow11_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow11_col5 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow12_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "            background-color:  #f7fbff;\n",
       "            color:  #000000;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow12_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow12_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow12_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow12_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow12_col5 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow13_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "            background-color:  #f7fbff;\n",
       "            color:  #000000;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow13_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow13_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow13_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow13_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow13_col5 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow14_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "            background-color:  #f7fbff;\n",
       "            color:  #000000;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow14_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow14_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow14_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow14_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow14_col5 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow15_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "            background-color:  #f7fbff;\n",
       "            color:  #000000;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow15_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow15_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow15_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow15_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow15_col5 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow16_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "            background-color:  #f7fbff;\n",
       "            color:  #000000;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow16_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow16_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow16_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow16_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow16_col5 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow17_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "            background-color:  #f7fbff;\n",
       "            color:  #000000;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow17_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow17_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow17_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow17_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow17_col5 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow18_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "            background-color:  #f7fbff;\n",
       "            color:  #000000;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow18_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow18_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow18_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow18_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow18_col5 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow19_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "            background-color:  #f7fbff;\n",
       "            color:  #000000;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow19_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow19_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow19_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow19_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow19_col5 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow20_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "            background-color:  #f7fbff;\n",
       "            color:  #000000;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow20_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow20_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow20_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow20_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow20_col5 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow21_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "            background-color:  #f7fbff;\n",
       "            color:  #000000;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow21_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow21_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow21_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow21_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow21_col5 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow22_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "            background-color:  #f7fbff;\n",
       "            color:  #000000;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow22_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow22_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow22_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow22_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow22_col5 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow23_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "            background-color:  #f7fbff;\n",
       "            color:  #000000;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow23_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow23_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow23_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow23_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow23_col5 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow24_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "            background-color:  #f7fbff;\n",
       "            color:  #000000;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow24_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow24_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow24_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow24_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow24_col5 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow25_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "            background-color:  #f7fbff;\n",
       "            color:  #000000;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow25_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow25_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow25_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow25_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow25_col5 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow26_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "            background-color:  #f7fbff;\n",
       "            color:  #000000;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow26_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow26_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow26_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow26_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow26_col5 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow27_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "            background-color:  #f7fbff;\n",
       "            color:  #000000;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow27_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow27_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow27_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow27_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow27_col5 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow28_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "            background-color:  #f7fbff;\n",
       "            color:  #000000;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow28_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow28_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow28_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow28_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow28_col5 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow29_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "            background-color:  #f7fbff;\n",
       "            color:  #000000;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow29_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow29_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow29_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow29_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow29_col5 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow30_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "            background-color:  #f7fbff;\n",
       "            color:  #000000;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow30_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow30_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow30_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow30_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow30_col5 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow31_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "            background-color:  #f7fbff;\n",
       "            color:  #000000;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow31_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow31_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow31_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow31_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow31_col5 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow32_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "            background-color:  #f7fbff;\n",
       "            color:  #000000;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow32_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow32_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow32_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow32_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow32_col5 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow33_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "            background-color:  #f7fbff;\n",
       "            color:  #000000;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow33_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow33_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow33_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow33_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow33_col5 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow34_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "            background-color:  #f7fbff;\n",
       "            color:  #000000;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow34_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow34_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow34_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow34_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow34_col5 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow35_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "            background-color:  #f7fbff;\n",
       "            color:  #000000;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow35_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow35_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow35_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow35_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow35_col5 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow36_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "            background-color:  #f7fbff;\n",
       "            color:  #000000;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow36_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow36_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow36_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow36_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow36_col5 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow37_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "            background-color:  #f7fbff;\n",
       "            color:  #000000;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow37_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow37_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow37_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow37_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow37_col5 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow38_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "            background-color:  #f7fbff;\n",
       "            color:  #000000;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow38_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow38_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow38_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow38_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow38_col5 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow39_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "            background-color:  #f7fbff;\n",
       "            color:  #000000;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow39_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow39_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow39_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow39_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow39_col5 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow40_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "            background-color:  #f7fbff;\n",
       "            color:  #000000;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow40_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow40_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow40_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow40_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow40_col5 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow41_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "            background-color:  #f7fbff;\n",
       "            color:  #000000;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow41_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow41_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow41_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow41_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow41_col5 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow42_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "            background-color:  #f7fbff;\n",
       "            color:  #000000;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow42_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow42_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow42_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow42_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow42_col5 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow43_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "            background-color:  #f7fbff;\n",
       "            color:  #000000;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow43_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow43_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow43_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow43_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow43_col5 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow44_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "            background-color:  #f7fbff;\n",
       "            color:  #000000;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow44_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow44_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow44_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow44_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow44_col5 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow45_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "            background-color:  #f7fbff;\n",
       "            color:  #000000;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow45_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow45_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow45_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow45_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow45_col5 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow46_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "            background-color:  #f7fbff;\n",
       "            color:  #000000;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow46_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow46_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow46_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow46_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow46_col5 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow47_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "            background-color:  #f7fbff;\n",
       "            color:  #000000;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow47_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow47_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow47_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow47_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow47_col5 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow48_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "            background-color:  #f7fbff;\n",
       "            color:  #000000;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow48_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow48_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow48_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow48_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow48_col5 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow49_col0 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  red;\n",
       "            background-color:  #f7fbff;\n",
       "            color:  #000000;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow49_col1 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow49_col2 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow49_col3 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow49_col4 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }    #T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow49_col5 {\n",
       "            background-color:  black;\n",
       "            color:  lawngreen;\n",
       "            border-color:  white;\n",
       "            color:  lawngreen;\n",
       "        }</style><table id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112f\" ><thead>    <tr>        <th class=\"blank level0\" ></th>        <th class=\"col_heading level0 col0\" >folder</th>        <th class=\"col_heading level0 col1\" >dialogueID</th>        <th class=\"col_heading level0 col2\" >date</th>        <th class=\"col_heading level0 col3\" >from</th>        <th class=\"col_heading level0 col4\" >to</th>        <th class=\"col_heading level0 col5\" >text</th>    </tr></thead><tbody>\n",
       "                <tr>\n",
       "                        <th id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112flevel0_row0\" class=\"row_heading level0 row0\" >1038274</th>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow0_col0\" class=\"data row0 col0\" >3</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow0_col1\" class=\"data row0 col1\" >91347.tsv</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow0_col2\" class=\"data row0 col2\" >2012-01-19T00:41:00.000Z</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow0_col3\" class=\"data row0 col3\" >natschil</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow0_col4\" class=\"data row0 col4\" >xiambax</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow0_col5\" class=\"data row0 col5\" >where?</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112flevel0_row1\" class=\"row_heading level0 row1\" >1038275</th>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow1_col0\" class=\"data row1 col0\" >3</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow1_col1\" class=\"data row1 col1\" >91347.tsv</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow1_col2\" class=\"data row1 col2\" >2012-01-19T00:41:00.000Z</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow1_col3\" class=\"data row1 col3\" >natschil</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow1_col4\" class=\"data row1 col4\" >xiambax</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow1_col5\" class=\"data row1 col5\" >i'm guessing gconf....</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112flevel0_row2\" class=\"row_heading level0 row2\" >1038276</th>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow2_col0\" class=\"data row2 col0\" >3</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow2_col1\" class=\"data row2 col1\" >272273.tsv</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow2_col2\" class=\"data row2 col2\" >2006-10-31T10:36:00.000Z</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow2_col3\" class=\"data row2 col3\" >ben_</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow2_col4\" class=\"data row2 col4\" >nan</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow2_col5\" class=\"data row2 col5\" >as in some of my RAM is bad?</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112flevel0_row3\" class=\"row_heading level0 row3\" >1038277</th>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow3_col0\" class=\"data row3 col0\" >3</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow3_col1\" class=\"data row3 col1\" >272273.tsv</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow3_col2\" class=\"data row3 col2\" >2006-10-31T10:38:00.000Z</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow3_col3\" class=\"data row3 col3\" >ben_</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow3_col4\" class=\"data row3 col4\" >nan</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow3_col5\" class=\"data row3 col5\" >i upgraded to edgy</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112flevel0_row4\" class=\"row_heading level0 row4\" >1038278</th>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow4_col0\" class=\"data row4 col0\" >3</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow4_col1\" class=\"data row4 col1\" >272273.tsv</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow4_col2\" class=\"data row4 col2\" >2006-10-31T10:40:00.000Z</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow4_col3\" class=\"data row4 col3\" >kbrosnan</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow4_col4\" class=\"data row4 col4\" >ben_</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow4_col5\" class=\"data row4 col5\" >are you using any plugins? about:plugins in the address bar that can also cause both to be crashy in the same way</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112flevel0_row5\" class=\"row_heading level0 row5\" >1038279</th>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow5_col0\" class=\"data row5 col0\" >3</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow5_col1\" class=\"data row5 col1\" >253128.tsv</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow5_col2\" class=\"data row5 col2\" >2009-02-17T06:31:00.000Z</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow5_col3\" class=\"data row5 col3\" >hlm</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow5_col4\" class=\"data row5 col4\" >nan</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow5_col5\" class=\"data row5 col5\" >ermmm I suppose im going to reinstall ubuntu for the 5th time in less than 48 hours :'(</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112flevel0_row6\" class=\"row_heading level0 row6\" >1038280</th>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow6_col0\" class=\"data row6 col0\" >3</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow6_col1\" class=\"data row6 col1\" >253128.tsv</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow6_col2\" class=\"data row6 col2\" >2009-02-17T06:32:00.000Z</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow6_col3\" class=\"data row6 col3\" >blouf</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow6_col4\" class=\"data row6 col4\" >hlm</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow6_col5\" class=\"data row6 col5\" >: have you tried to customize usplash ?</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112flevel0_row7\" class=\"row_heading level0 row7\" >1038281</th>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow7_col0\" class=\"data row7 col0\" >3</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow7_col1\" class=\"data row7 col1\" >253128.tsv</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow7_col2\" class=\"data row7 col2\" >2009-02-17T06:34:00.000Z</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow7_col3\" class=\"data row7 col3\" >blouf</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow7_col4\" class=\"data row7 col4\" >hlm</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow7_col5\" class=\"data row7 col5\" >: have you tried to customize usplash ?</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112flevel0_row8\" class=\"row_heading level0 row8\" >1038282</th>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow8_col0\" class=\"data row8 col0\" >3</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow8_col1\" class=\"data row8 col1\" >84896.tsv</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow8_col2\" class=\"data row8 col2\" >2009-01-13T18:30:00.000Z</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow8_col3\" class=\"data row8 col3\" >EdgeX-</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow8_col4\" class=\"data row8 col4\" >nan</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow8_col5\" class=\"data row8 col5\" >Hello; What program would you recommend for me as far as video editing goes?</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112flevel0_row9\" class=\"row_heading level0 row9\" >1038283</th>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow9_col0\" class=\"data row9 col0\" >3</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow9_col1\" class=\"data row9 col1\" >84896.tsv</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow9_col2\" class=\"data row9 col2\" >2009-01-13T18:31:00.000Z</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow9_col3\" class=\"data row9 col3\" >EdgeX-</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow9_col4\" class=\"data row9 col4\" >nan</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow9_col5\" class=\"data row9 col5\" >thanks jrib :)</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112flevel0_row10\" class=\"row_heading level0 row10\" >1038284</th>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow10_col0\" class=\"data row10 col0\" >3</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow10_col1\" class=\"data row10 col1\" >84896.tsv</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow10_col2\" class=\"data row10 col2\" >2009-01-13T18:31:00.000Z</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow10_col3\" class=\"data row10 col3\" >Deckard_BR26354</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow10_col4\" class=\"data row10 col4\" >EdgeX-</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow10_col5\" class=\"data row10 col5\" >avidemux (GTK+)</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112flevel0_row11\" class=\"row_heading level0 row11\" >1038285</th>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow11_col0\" class=\"data row11 col0\" >3</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow11_col1\" class=\"data row11 col1\" >231761.tsv</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow11_col2\" class=\"data row11 col2\" >2006-08-23T06:35:00.000Z</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow11_col3\" class=\"data row11 col3\" >Jbirk</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow11_col4\" class=\"data row11 col4\" >nan</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow11_col5\" class=\"data row11 col5\" >Hi</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112flevel0_row12\" class=\"row_heading level0 row12\" >1038286</th>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow12_col0\" class=\"data row12 col0\" >3</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow12_col1\" class=\"data row12 col1\" >231761.tsv</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow12_col2\" class=\"data row12 col2\" >2006-08-23T06:36:00.000Z</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow12_col3\" class=\"data row12 col3\" >Jbirk</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow12_col4\" class=\"data row12 col4\" >nan</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow12_col5\" class=\"data row12 col5\" >How do I install and use wine?</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112flevel0_row13\" class=\"row_heading level0 row13\" >1038287</th>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow13_col0\" class=\"data row13 col0\" >3</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow13_col1\" class=\"data row13 col1\" >231761.tsv</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow13_col2\" class=\"data row13 col2\" >2006-08-23T06:36:00.000Z</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow13_col3\" class=\"data row13 col3\" >Zambezi</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow13_col4\" class=\"data row13 col4\" >Jbirk</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow13_col5\" class=\"data row13 col5\" >sudo apt-get install wine is the beginning.</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112flevel0_row14\" class=\"row_heading level0 row14\" >1038288</th>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow14_col0\" class=\"data row14 col0\" >3</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow14_col1\" class=\"data row14 col1\" >204548.tsv</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow14_col2\" class=\"data row14 col2\" >2006-10-19T08:46:00.000Z</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow14_col3\" class=\"data row14 col3\" >nikko82</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow14_col4\" class=\"data row14 col4\" >nan</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow14_col5\" class=\"data row14 col5\" >does anyone know what can cause a 'segmentation fault' when a use depmod?</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112flevel0_row15\" class=\"row_heading level0 row15\" >1038289</th>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow15_col0\" class=\"data row15 col0\" >3</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow15_col1\" class=\"data row15 col1\" >204548.tsv</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow15_col2\" class=\"data row15 col2\" >2006-10-19T08:46:00.000Z</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow15_col3\" class=\"data row15 col3\" >frogzoo</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow15_col4\" class=\"data row15 col4\" >nikko82</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow15_col5\" class=\"data row15 col5\" >strace</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112flevel0_row16\" class=\"row_heading level0 row16\" >1038290</th>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow16_col0\" class=\"data row16 col0\" >3</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow16_col1\" class=\"data row16 col1\" >204548.tsv</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow16_col2\" class=\"data row16 col2\" >2006-10-19T08:47:00.000Z</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow16_col3\" class=\"data row16 col3\" >nikko82</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow16_col4\" class=\"data row16 col4\" >frogzoo</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow16_col5\" class=\"data row16 col5\" >thanks. how do i use it?</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112flevel0_row17\" class=\"row_heading level0 row17\" >1038291</th>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow17_col0\" class=\"data row17 col0\" >3</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow17_col1\" class=\"data row17 col1\" >107510.tsv</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow17_col2\" class=\"data row17 col2\" >2008-06-23T08:50:00.000Z</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow17_col3\" class=\"data row17 col3\" >torias</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow17_col4\" class=\"data row17 col4\" >nan</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow17_col5\" class=\"data row17 col5\" >Hello, I tried to go into the advanced desktop settings by typing sudo apt-get install compizconfig-settings-manager, and it pulled up sudo password for torias:, i couldnt enter anything and didnt even set up a sudo password, can someone please help me be able to configure the advanced desktop settings</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112flevel0_row18\" class=\"row_heading level0 row18\" >1038292</th>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow18_col0\" class=\"data row18 col0\" >3</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow18_col1\" class=\"data row18 col1\" >107510.tsv</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow18_col2\" class=\"data row18 col2\" >2008-06-23T08:51:00.000Z</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow18_col3\" class=\"data row18 col3\" >torias</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow18_col4\" class=\"data row18 col4\" >nan</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow18_col5\" class=\"data row18 col5\" >i typed this in the terminal</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112flevel0_row19\" class=\"row_heading level0 row19\" >1038293</th>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow19_col0\" class=\"data row19 col0\" >3</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow19_col1\" class=\"data row19 col1\" >107510.tsv</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow19_col2\" class=\"data row19 col2\" >2008-06-23T08:51:00.000Z</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow19_col3\" class=\"data row19 col3\" >thejaja</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow19_col4\" class=\"data row19 col4\" >torias</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow19_col5\" class=\"data row19 col5\" >use the user pass</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112flevel0_row20\" class=\"row_heading level0 row20\" >1038294</th>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow20_col0\" class=\"data row20 col0\" >3</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow20_col1\" class=\"data row20 col1\" >79088.tsv</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow20_col2\" class=\"data row20 col2\" >2011-10-14T04:52:00.000Z</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow20_col3\" class=\"data row20 col3\" >rahul</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow20_col4\" class=\"data row20 col4\" >nan</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow20_col5\" class=\"data row20 col5\" >They ubuntu alternate gives more support to h/w. What kinda support? Please.</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112flevel0_row21\" class=\"row_heading level0 row21\" >1038295</th>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow21_col0\" class=\"data row21 col0\" >3</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow21_col1\" class=\"data row21 col1\" >79088.tsv</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow21_col2\" class=\"data row21 col2\" >2011-10-14T04:52:00.000Z</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow21_col3\" class=\"data row21 col3\" >rahul</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow21_col4\" class=\"data row21 col4\" >nan</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow21_col5\" class=\"data row21 col5\" >They say ubuntu alternate gives more support to h/w. What kinda support? Please.</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112flevel0_row22\" class=\"row_heading level0 row22\" >1038296</th>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow22_col0\" class=\"data row22 col0\" >3</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow22_col1\" class=\"data row22 col1\" >79088.tsv</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow22_col2\" class=\"data row22 col2\" >2011-10-14T04:53:00.000Z</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow22_col3\" class=\"data row22 col3\" >xangua</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow22_col4\" class=\"data row22 col4\" >rahul</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow22_col5\" class=\"data row22 col5\" >is ubutnu, the same kernel, support is the same</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112flevel0_row23\" class=\"row_heading level0 row23\" >1038297</th>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow23_col0\" class=\"data row23 col0\" >3</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow23_col1\" class=\"data row23 col1\" >207077.tsv</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow23_col2\" class=\"data row23 col2\" >2006-10-12T12:51:00.000Z</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow23_col3\" class=\"data row23 col3\" >seamus7</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow23_col4\" class=\"data row23 col4\" >nan</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow23_col5\" class=\"data row23 col5\" >hi all ... just curious ... wondering if many here have kubuntu-desktop intalled as a second gui and wondering what you think might be a strong reason for having it installed in addition to Gnome? I just installed it to check it out but wonder if there's a 'killer app' that perhaps I can only use in Kubuntu that I couldn't use in Ubuntu like I do with Amarok.</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112flevel0_row24\" class=\"row_heading level0 row24\" >1038298</th>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow24_col0\" class=\"data row24 col0\" >3</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow24_col1\" class=\"data row24 col1\" >207077.tsv</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow24_col2\" class=\"data row24 col2\" >2006-10-12T12:52:00.000Z</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow24_col3\" class=\"data row24 col3\" >Rico-</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow24_col4\" class=\"data row24 col4\" >seamus7</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow24_col5\" class=\"data row24 col5\" >all kde apps run on gnome, just with differing amounts of kde installed</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112flevel0_row25\" class=\"row_heading level0 row25\" >1038299</th>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow25_col0\" class=\"data row25 col0\" >3</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow25_col1\" class=\"data row25 col1\" >207077.tsv</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow25_col2\" class=\"data row25 col2\" >2006-10-12T12:53:00.000Z</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow25_col3\" class=\"data row25 col3\" >seamus7</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow25_col4\" class=\"data row25 col4\" >Rico-</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow25_col5\" class=\"data row25 col5\" >ahhh ok I didn't know you could use all KDE apps on Gnome. Thx.</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112flevel0_row26\" class=\"row_heading level0 row26\" >1038300</th>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow26_col0\" class=\"data row26 col0\" >3</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow26_col1\" class=\"data row26 col1\" >224282.tsv</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow26_col2\" class=\"data row26 col2\" >2010-04-20T16:30:00.000Z</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow26_col3\" class=\"data row26 col3\" >alexs</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow26_col4\" class=\"data row26 col4\" >nan</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow26_col5\" class=\"data row26 col5\" >Anyone know where the edgy repo in old-releases has gone?</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112flevel0_row27\" class=\"row_heading level0 row27\" >1038301</th>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow27_col0\" class=\"data row27 col0\" >3</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow27_col1\" class=\"data row27 col1\" >224282.tsv</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow27_col2\" class=\"data row27 col2\" >2010-04-20T16:30:00.000Z</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow27_col3\" class=\"data row27 col3\" >alexs</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow27_col4\" class=\"data row27 col4\" >nan</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow27_col5\" class=\"data row27 col5\" >http://old-releases.ubuntu.com/releases/edgy/ 404s</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112flevel0_row28\" class=\"row_heading level0 row28\" >1038302</th>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow28_col0\" class=\"data row28 col0\" >3</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow28_col1\" class=\"data row28 col1\" >224282.tsv</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow28_col2\" class=\"data row28 col2\" >2010-04-20T16:31:00.000Z</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow28_col3\" class=\"data row28 col3\" >jrib</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow28_col4\" class=\"data row28 col4\" >alexs</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow28_col5\" class=\"data row28 col5\" >edgy has not been supported for a long time</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112flevel0_row29\" class=\"row_heading level0 row29\" >1038303</th>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow29_col0\" class=\"data row29 col0\" >3</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow29_col1\" class=\"data row29 col1\" >299062.tsv</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow29_col2\" class=\"data row29 col2\" >2005-09-10T10:36:00.000Z</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow29_col3\" class=\"data row29 col3\" >TrekCycling</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow29_col4\" class=\"data row29 col4\" >nan</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow29_col5\" class=\"data row29 col5\" >what kernel is the next ubuntu release supposed to use?</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112flevel0_row30\" class=\"row_heading level0 row30\" >1038304</th>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow30_col0\" class=\"data row30 col0\" >3</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow30_col1\" class=\"data row30 col1\" >299062.tsv</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow30_col2\" class=\"data row30 col2\" >2005-09-10T10:38:00.000Z</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow30_col3\" class=\"data row30 col3\" >TrekCycling</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow30_col4\" class=\"data row30 col4\" >nan</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow30_col5\" class=\"data row30 col5\" >does anyone really use Breezy? or not until oct.</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112flevel0_row31\" class=\"row_heading level0 row31\" >1038305</th>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow31_col0\" class=\"data row31 col0\" >3</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow31_col1\" class=\"data row31 col1\" >299062.tsv</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow31_col2\" class=\"data row31 col2\" >2005-09-10T10:38:00.000Z</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow31_col3\" class=\"data row31 col3\" >CarlFK</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow31_col4\" class=\"data row31 col4\" >TrekCycling</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow31_col5\" class=\"data row31 col5\" >- you can use it as long as you don't expect it to work</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112flevel0_row32\" class=\"row_heading level0 row32\" >1038306</th>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow32_col0\" class=\"data row32 col0\" >3</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow32_col1\" class=\"data row32 col1\" >195440.tsv</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow32_col2\" class=\"data row32 col2\" >2007-01-10T10:33:00.000Z</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow32_col3\" class=\"data row32 col3\" >puff</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow32_col4\" class=\"data row32 col4\" >nan</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow32_col5\" class=\"data row32 col5\" >Anybody got an ipod? I'm seeing some wacky behavior with gtkpod and I'm trying to figure out what I'm doing wrong.</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112flevel0_row33\" class=\"row_heading level0 row33\" >1038307</th>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow33_col0\" class=\"data row33 col0\" >3</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow33_col1\" class=\"data row33 col1\" >195440.tsv</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow33_col2\" class=\"data row33 col2\" >2007-01-10T10:33:00.000Z</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow33_col3\" class=\"data row33 col3\" >riotkittie</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow33_col4\" class=\"data row33 col4\" >puff</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow33_col5\" class=\"data row33 col5\" > i dont have an ipod but if you mail me yours, i'll look into your issue. heh heh.</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112flevel0_row34\" class=\"row_heading level0 row34\" >1038308</th>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow34_col0\" class=\"data row34 col0\" >3</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow34_col1\" class=\"data row34 col1\" >195440.tsv</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow34_col2\" class=\"data row34 col2\" >2007-01-10T10:34:00.000Z</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow34_col3\" class=\"data row34 col3\" >puff</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow34_col4\" class=\"data row34 col4\" >riotkittie</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow34_col5\" class=\"data row34 col5\" >Gee, thanks :-).</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112flevel0_row35\" class=\"row_heading level0 row35\" >1038309</th>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow35_col0\" class=\"data row35 col0\" >3</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow35_col1\" class=\"data row35 col1\" >176570.tsv</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow35_col2\" class=\"data row35 col2\" >2007-03-20T10:54:00.000Z</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow35_col3\" class=\"data row35 col3\" >fernandolima</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow35_col4\" class=\"data row35 col4\" >nan</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow35_col5\" class=\"data row35 col5\" >hello</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112flevel0_row36\" class=\"row_heading level0 row36\" >1038310</th>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow36_col0\" class=\"data row36 col0\" >3</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow36_col1\" class=\"data row36 col1\" >176570.tsv</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow36_col2\" class=\"data row36 col2\" >2007-03-20T10:54:00.000Z</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow36_col3\" class=\"data row36 col3\" >fernandolima</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow36_col4\" class=\"data row36 col4\" >nan</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow36_col5\" class=\"data row36 col5\" >somebody can help me</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112flevel0_row37\" class=\"row_heading level0 row37\" >1038311</th>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow37_col0\" class=\"data row37 col0\" >3</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow37_col1\" class=\"data row37 col1\" >176570.tsv</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow37_col2\" class=\"data row37 col2\" >2007-03-20T10:55:00.000Z</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow37_col3\" class=\"data row37 col3\" >mjbrooks</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow37_col4\" class=\"data row37 col4\" >fernandolima</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow37_col5\" class=\"data row37 col5\" >depends on what you need</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112flevel0_row38\" class=\"row_heading level0 row38\" >1038312</th>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow38_col0\" class=\"data row38 col0\" >3</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow38_col1\" class=\"data row38 col1\" >72561.tsv</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow38_col2\" class=\"data row38 col2\" >2010-05-14T04:44:00.000Z</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow38_col3\" class=\"data row38 col3\" >lokpest_</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow38_col4\" class=\"data row38 col4\" >nan</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow38_col5\" class=\"data row38 col5\" >How do i connect to internet from the console where I dont have access to any graphic network manager?</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112flevel0_row39\" class=\"row_heading level0 row39\" >1038313</th>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow39_col0\" class=\"data row39 col0\" >3</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow39_col1\" class=\"data row39 col1\" >72561.tsv</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow39_col2\" class=\"data row39 col2\" >2010-05-14T04:45:00.000Z</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow39_col3\" class=\"data row39 col3\" >skierpage</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow39_col4\" class=\"data row39 col4\" >lokpest_</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow39_col5\" class=\"data row39 col5\" >what did nm-tool print out?</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112flevel0_row40\" class=\"row_heading level0 row40\" >1038314</th>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow40_col0\" class=\"data row40 col0\" >3</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow40_col1\" class=\"data row40 col1\" >72561.tsv</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow40_col2\" class=\"data row40 col2\" >2010-05-14T04:52:00.000Z</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow40_col3\" class=\"data row40 col3\" >skierpage</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow40_col4\" class=\"data row40 col4\" >lokpest_</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow40_col5\" class=\"data row40 col5\" >  I'm not sure if NetworkManager is started at desktop login.  Try `ps alx | grep -i net` and see if anything's running.  You can restart it with  maybe `restart network-manager` ??</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112flevel0_row41\" class=\"row_heading level0 row41\" >1038315</th>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow41_col0\" class=\"data row41 col0\" >3</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow41_col1\" class=\"data row41 col1\" >19585.tsv</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow41_col2\" class=\"data row41 col2\" >2011-04-04T21:55:00.000Z</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow41_col3\" class=\"data row41 col3\" >CarlFK</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow41_col4\" class=\"data row41 col4\" >benzaldehyde</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow41_col5\" class=\"data row41 col5\" >S-video can do sound?</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112flevel0_row42\" class=\"row_heading level0 row42\" >1038316</th>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow42_col0\" class=\"data row42 col0\" >3</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow42_col1\" class=\"data row42 col1\" >19585.tsv</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow42_col2\" class=\"data row42 col2\" >2011-04-04T21:55:00.000Z</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow42_col3\" class=\"data row42 col3\" >benzaldehyde</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow42_col4\" class=\"data row42 col4\" >CarlFK</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow42_col5\" class=\"data row42 col5\" >well maybe not sound, i don't remember</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112flevel0_row43\" class=\"row_heading level0 row43\" >1038317</th>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow43_col0\" class=\"data row43 col0\" >3</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow43_col1\" class=\"data row43 col1\" >19585.tsv</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow43_col2\" class=\"data row43 col2\" >2011-04-04T21:56:00.000Z</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow43_col3\" class=\"data row43 col3\" >CarlFK</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow43_col4\" class=\"data row43 col4\" >benzaldehyde</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow43_col5\" class=\"data row43 col5\" >'S-Video carries standard definition video (typically at 480i or 576i resolution), but does not carry audio on the same cable.'  http://en.wikipedia.org/wiki/Svideo</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112flevel0_row44\" class=\"row_heading level0 row44\" >1038318</th>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow44_col0\" class=\"data row44 col0\" >3</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow44_col1\" class=\"data row44 col1\" >51506.tsv</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow44_col2\" class=\"data row44 col2\" >2012-01-31T10:55:00.000Z</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow44_col3\" class=\"data row44 col3\" >DJ</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow44_col4\" class=\"data row44 col4\" >nan</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow44_col5\" class=\"data row44 col5\" >?</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112flevel0_row45\" class=\"row_heading level0 row45\" >1038319</th>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow45_col0\" class=\"data row45 col0\" >3</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow45_col1\" class=\"data row45 col1\" >51506.tsv</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow45_col2\" class=\"data row45 col2\" >2012-01-31T10:56:00.000Z</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow45_col3\" class=\"data row45 col3\" >DJ</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow45_col4\" class=\"data row45 col4\" >nan</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow45_col5\" class=\"data row45 col5\" >anyone on?</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112flevel0_row46\" class=\"row_heading level0 row46\" >1038320</th>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow46_col0\" class=\"data row46 col0\" >3</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow46_col1\" class=\"data row46 col1\" >51506.tsv</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow46_col2\" class=\"data row46 col2\" >2012-01-31T10:56:00.000Z</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow46_col3\" class=\"data row46 col3\" >aeon-ltd</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow46_col4\" class=\"data row46 col4\" >DJ</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow46_col5\" class=\"data row46 col5\" >yes</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112flevel0_row47\" class=\"row_heading level0 row47\" >1038321</th>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow47_col0\" class=\"data row47 col0\" >3</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow47_col1\" class=\"data row47 col1\" >99669.tsv</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow47_col2\" class=\"data row47 col2\" >2008-11-16T20:11:00.000Z</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow47_col3\" class=\"data row47 col3\" >KR-data</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow47_col4\" class=\"data row47 col4\" >nan</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow47_col5\" class=\"data row47 col5\" >can I get a pastebin of someones menu.lst with a working dualboot with XP?</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112flevel0_row48\" class=\"row_heading level0 row48\" >1038322</th>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow48_col0\" class=\"data row48 col0\" >3</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow48_col1\" class=\"data row48 col1\" >99669.tsv</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow48_col2\" class=\"data row48 col2\" >2008-11-16T20:12:00.000Z</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow48_col3\" class=\"data row48 col3\" >outbackwifi</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow48_col4\" class=\"data row48 col4\" >KR-data</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow48_col5\" class=\"data row48 col5\" >http://pastebin.com/fe921690</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112flevel0_row49\" class=\"row_heading level0 row49\" >1038323</th>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow49_col0\" class=\"data row49 col0\" >3</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow49_col1\" class=\"data row49 col1\" >99669.tsv</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow49_col2\" class=\"data row49 col2\" >2008-11-16T20:13:00.000Z</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow49_col3\" class=\"data row49 col3\" >KR-data</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow49_col4\" class=\"data row49 col4\" >outbackwifi</td>\n",
       "                        <td id=\"T_7aa6887c_cf1e_11ea_8941_acbc32ce112frow49_col5\" class=\"data row49 col5\" >thanks</td>\n",
       "            </tr>\n",
       "    </tbody></table>"
      ],
      "text/plain": [
       "<pandas.io.formats.style.Styler at 0x19536fdb50>"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "show(ubuntu_df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 1038324 entries, 0 to 1038323\n",
      "Data columns (total 6 columns):\n",
      " #   Column      Non-Null Count    Dtype \n",
      "---  ------      --------------    ----- \n",
      " 0   folder      1038324 non-null  int64 \n",
      " 1   dialogueID  1038324 non-null  object\n",
      " 2   date        1038324 non-null  object\n",
      " 3   from        1038311 non-null  object\n",
      " 4   to          566035 non-null   object\n",
      " 5   text        1038237 non-null  object\n",
      "dtypes: int64(1), object(5)\n",
      "memory usage: 47.5+ MB\n"
     ]
    }
   ],
   "source": [
    "ubuntu_df.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0          Hello folks, please help me a bit with the fol...\n",
       "1          Did I choose a bad channel? I ask because you ...\n",
       "2          the second sentence is better english   and we...\n",
       "3                                               Sock Puppe?t\n",
       "4                                                       WTF?\n",
       "                                 ...                        \n",
       "1038319                                           anyone on?\n",
       "1038320                                                  yes\n",
       "1038321    can I get a pastebin of someones menu.lst with...\n",
       "1038322                         http://pastebin.com/fe921690\n",
       "1038323                                               thanks\n",
       "Name: text, Length: 1038324, dtype: object"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Getting just the text\n",
    "text = pd.Series(ubuntu_df['text']).astype(str)\n",
    "text"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [],
   "source": [
    "# End to end tokenizer function\n",
    "\n",
    "def ubuntu_tokenizer(data: lambda _data: isinstance(_data, pd.Series)) -> 'Cleaned Pandas Series':\n",
    "    ''' I am making my own end-to-end tokenizer function for preprocessing that accepts\n",
    "    a Pandas Series as input and outputs a preprocessed Pandas Series'''\n",
    "    \n",
    "    # Making sure input is a series (these are two ways of doing the same thing)\n",
    "    assert isinstance(data,pd.Series), 'Input must be a Pandas Series'\n",
    "    \n",
    "    # 1. Converting all to lower case\n",
    "    data = data.str.lower()\n",
    "    \n",
    "    print(f'1. Original shape of data is {data.shape}')\n",
    "    \n",
    "    \n",
    "    # 2. Tokenizing with NLTK's TweetTokenizer. This limits repeated characters to \n",
    "    # three with the reduce lens paramater and strips all the @'s. It also\n",
    "    # splits it into 1-gram tokens\n",
    "    tknzr = TweetTokenizer(strip_handles = True, reduce_len = True)\n",
    "    # Using progress apply to show the progress bar\n",
    "    data = data.progress_apply(tknzr.tokenize)\n",
    "    \n",
    "    print(f'2. Tokenized, removed handles, and reduced the length of repeated characters.\\\n",
    "          \\n Shape is still {data.shape}. \\n \\n Peek: \\n {data.head()}')\n",
    "    visualize_lengths(data, 'Length of Tokens after Step 2')\n",
    "    \n",
    "    # 3. Removing the punctuation\n",
    "    \n",
    "    data = data.progress_apply(remove_from_list, stuff_to_remove = puncts)\n",
    "    print(f'3. Removed empty tokens and punctuation. Shape is still {data.shape}. \\n \\n Peek: \\n {data.head()}')\n",
    "    visualize_lengths(data, 'Length of Tokens after Step 9')\n",
    "    \n",
    "    # 3.1. Removing links\n",
    "    \n",
    "    data = data.progress_apply(remove_links)\n",
    "    \n",
    "    print(f'3. Removed the links. Shape is still {data.shape}. \\n \\n Peek: \\n {data.head()}')\n",
    "    visualize_lengths(data, 'Length of Tokens after Step 3')\n",
    "\n",
    "    # 4. Checking for and correcting mispellings\n",
    "    spell = SpellChecker()\n",
    "    \n",
    "    # Skipping this step first because it takes way too long\n",
    "#     data = data.progress_apply(correct_spellings)\n",
    "    \n",
    "    print(f'4. Applied automispelling corrections. Shape is still {data.shape}. \\n \\n Peek: \\n {data.head()}')\n",
    "    \n",
    "    # 5. Removing the stop words, utilizing the same remove_from_list function defined above\n",
    "    \n",
    "    data = data.progress_apply(remove_from_list, stuff_to_remove = stop_words)\n",
    "    \n",
    "    print(f'5. Removed the stop words. Shape is still {data.shape}. \\n \\n Peek: \\n {data.head()}')\n",
    "    visualize_lengths(data, 'Length of Tokens after Step 4')\n",
    "    \n",
    "    # 6. Expanding contractions and colloquial language\n",
    "    \n",
    "    data = data.progress_apply(replace_from_dict, dic = contraction_colloq_dict)\n",
    "    print(f'6. Expanded contractions into extra tokens. Shape is still {data.shape}. \\n \\n Peek: \\n {data.head()}')\n",
    "    visualize_lengths(data, 'Length of Tokens after Step 6')\n",
    "\n",
    "def ubuntu_tokenizer_2(data: lambda _data: isinstance(_data, pd.Series)):\n",
    "    \n",
    "    # 7. Removing non-english Tweets with spaCy\n",
    "\n",
    "    data = data[only_english(data)]\n",
    "    print(f'7. Remove all non-english Tweets. Shape is now {data.shape}. Clearly less than before! \\n \\n Peek: \\n {data.head()}')\n",
    "    \n",
    "    # 8. Lemmatization\n",
    "\n",
    "    data = data.progress_apply(lemmatize_list)\n",
    "    print(f'8. Lemmatized the tokens. Shape is still {data.shape}. \\n \\n Peek: \\n {data.head()}')\n",
    "    visualize_lengths(data, 'Length of Tokens after Step 8')   \n",
    "   \n",
    "    # Removing again to make sure I get everything\n",
    "    data = data.progress_apply(remove_from_list, stuff_to_remove = puncts)\n",
    "\n",
    "    # 9.1. Removing emojis -- (UPDATE) in a way that preserves Series indexes\n",
    "    \n",
    "    unique_emojis = [i[0] for i in emoji.UNICODE_EMOJI]\n",
    "    data = data.apply(remove_from_list, stuff_to_remove = unique_emojis)\n",
    "    \n",
    "    # 9.2. Removing numbers -- (UPDATE) also in a way that preserves Series indexes\n",
    "    # Using nested list comprehension\n",
    "    data = data.progress_apply(clean_numbers)\n",
    "    \n",
    "    print(f'9. Removed emojis and numbers. Shape is still {data.shape}. \\n \\n Peek: \\n {data.head()}')\n",
    "    visualize_lengths(data, 'Length of Tokens after Step 9')   \n",
    "    \n",
    "    # 10. Limiting length of Tweet\n",
    "    max_tokens = 50\n",
    "    min_tokens = 5\n",
    "    data = data.progress_apply(limit_length, min_tokens = min_tokens, max_tokens = max_tokens)\n",
    "    # Dropping all nan values, which are the token limits that didn't meet the thresholding requirements\n",
    "    data = data.dropna()\n",
    "    print(f'10. Limited each tweet to a max of {max_tokens} tokens and a min of {min_tokens} tokens. Shape is now {data.shape}. \\n \\n Peek: \\n {data.head()}')\n",
    "    visualize_lengths(data, 'Length of Tokens after Step 10')\n",
    "    \n",
    "    return data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1. Original shape of data is (1038324,)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "b1757e79205d46ccae61a000964c9d32",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=1038324.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Applying first part of preprocessing pipeline\n",
    "ubuntu_tokenizer(text)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['Color',\n",
       " 'ColorBar',\n",
       " 'ColumnDataSource',\n",
       " 'CountVectorizer',\n",
       " 'Counter',\n",
       " 'Emoji',\n",
       " 'In',\n",
       " 'Out',\n",
       " 'SpellChecker',\n",
       " 'TweetTokenizer',\n",
       " 'WordNetLemmatizer',\n",
       " '_',\n",
       " '_10',\n",
       " '_14',\n",
       " '_15',\n",
       " '_16',\n",
       " '_18',\n",
       " '_8',\n",
       " '__',\n",
       " '___',\n",
       " '__builtin__',\n",
       " '__builtins__',\n",
       " '__doc__',\n",
       " '__loader__',\n",
       " '__name__',\n",
       " '__package__',\n",
       " '__spec__',\n",
       " '_dh',\n",
       " '_exit_code',\n",
       " '_i',\n",
       " '_i1',\n",
       " '_i10',\n",
       " '_i11',\n",
       " '_i12',\n",
       " '_i13',\n",
       " '_i14',\n",
       " '_i15',\n",
       " '_i16',\n",
       " '_i17',\n",
       " '_i18',\n",
       " '_i19',\n",
       " '_i2',\n",
       " '_i20',\n",
       " '_i21',\n",
       " '_i22',\n",
       " '_i23',\n",
       " '_i24',\n",
       " '_i3',\n",
       " '_i4',\n",
       " '_i5',\n",
       " '_i6',\n",
       " '_i7',\n",
       " '_i8',\n",
       " '_i9',\n",
       " '_ih',\n",
       " '_ii',\n",
       " '_iii',\n",
       " '_oh',\n",
       " 'all_data',\n",
       " 'apple_data',\n",
       " 'authors_count',\n",
       " 'ax',\n",
       " 'c',\n",
       " 'casual_tokenize',\n",
       " 'clean_numbers',\n",
       " 'contraction_colloq_dict',\n",
       " 'correct_spellings',\n",
       " 'count',\n",
       " 'df',\n",
       " 'emoji',\n",
       " 'en_core_web_sm',\n",
       " 'exit',\n",
       " 'extract_emojis',\n",
       " 'figure',\n",
       " 'get_ipython',\n",
       " 'get_wordnet_pos',\n",
       " 'inbound',\n",
       " 'language_detector',\n",
       " 'lemmatize_list',\n",
       " 'lemmatizer',\n",
       " 'limit_length',\n",
       " 'my_tokenizer',\n",
       " 'nlp_cld',\n",
       " 'nltk',\n",
       " 'np',\n",
       " 'only_english',\n",
       " 'outbound',\n",
       " 'output_file',\n",
       " 'pd',\n",
       " 'plt',\n",
       " 'punct_base',\n",
       " 'puncts',\n",
       " 'quit',\n",
       " 're',\n",
       " 'remove_from_list',\n",
       " 'remove_links',\n",
       " 'replace_from_dict',\n",
       " 'show',\n",
       " 'sns',\n",
       " 'spacy',\n",
       " 'spacy_cld',\n",
       " 'st',\n",
       " 'stop_words',\n",
       " 'stopwords',\n",
       " 'text',\n",
       " 'top10_bagofwords',\n",
       " 'tqdm',\n",
       " 'ubuntu_df',\n",
       " 'validate',\n",
       " 'visualize_lengths',\n",
       " 'word_tokenize',\n",
       " 'wordnet',\n",
       " 'x',\n",
       " 'yaml']"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dir()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This is a better place to get intents like gratitude or goodbye."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Relational Services in Customer Service (RSICS)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "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>Annotator A ID</th>\n",
       "      <th>Annotator B ID</th>\n",
       "      <th>Parition ID</th>\n",
       "      <th>Corpora ID</th>\n",
       "      <th>Sentence ID</th>\n",
       "      <th>Text</th>\n",
       "      <th>Annotator A Text</th>\n",
       "      <th>Annotator B Text</th>\n",
       "      <th>Length</th>\n",
       "      <th>Error</th>\n",
       "      <th>Alignment Score</th>\n",
       "      <th>Agreement</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>7.000000</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>7506.0</td>\n",
       "      <td>what is the price for round trip between toron...</td>\n",
       "      <td>what is the price for round trip between toron...</td>\n",
       "      <td>what is the price for round trip between toron...</td>\n",
       "      <td>61.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>7.000000</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>7507.0</td>\n",
       "      <td>trying to find a good deal and good timing to ...</td>\n",
       "      <td>trying to find a good deal and good timing to ...</td>\n",
       "      <td>[trying to find a ]good deal and good timing t...</td>\n",
       "      <td>81.0</td>\n",
       "      <td>17.0</td>\n",
       "      <td>0.790123</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>7.000000</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>7509.0</td>\n",
       "      <td>trying to plan route penn ny to montgomery ny....</td>\n",
       "      <td>trying to plan route penn ny to montgomery ny....</td>\n",
       "      <td>trying to plan route penn ny to montgomery ny....</td>\n",
       "      <td>63.0</td>\n",
       "      <td>16.0</td>\n",
       "      <td>0.746032</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>7.000000</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>7510.0</td>\n",
       "      <td>Hi does the 54 Vermonter train go from Newark,...</td>\n",
       "      <td>[Hi ]does the 54 Vermonter train go from Newar...</td>\n",
       "      <td>[Hi ]does the 54 Vermonter train go from Newar...</td>\n",
       "      <td>87.0</td>\n",
       "      <td>10.0</td>\n",
       "      <td>0.885057</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1</td>\n",
       "      <td>7.000000</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>7511.0</td>\n",
       "      <td>is there a train service from los angeles ca t...</td>\n",
       "      <td>is there a train service from los angeles ca t...</td>\n",
       "      <td>is there a train service from los angeles ca t...</td>\n",
       "      <td>64.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.0</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",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2658</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2659</th>\n",
       "      <td>Average Length</td>\n",
       "      <td>97.922122</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2660</th>\n",
       "      <td>Average Error</td>\n",
       "      <td>0.076430</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2661</th>\n",
       "      <td>Average Alignment</td>\n",
       "      <td>0.923570</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2662</th>\n",
       "      <td>Average Agreement</td>\n",
       "      <td>0.787434</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>2663 rows × 12 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "         Annotator A ID  Annotator B ID  Parition ID  Corpora ID  Sentence ID  \\\n",
       "0                     1        7.000000          2.0         2.0       7506.0   \n",
       "1                     1        7.000000          2.0         2.0       7507.0   \n",
       "2                     1        7.000000          2.0         2.0       7509.0   \n",
       "3                     1        7.000000          2.0         2.0       7510.0   \n",
       "4                     1        7.000000          2.0         2.0       7511.0   \n",
       "...                 ...             ...          ...         ...          ...   \n",
       "2658                NaN             NaN          NaN         NaN          NaN   \n",
       "2659     Average Length       97.922122          NaN         NaN          NaN   \n",
       "2660      Average Error        0.076430          NaN         NaN          NaN   \n",
       "2661  Average Alignment        0.923570          NaN         NaN          NaN   \n",
       "2662  Average Agreement        0.787434          NaN         NaN          NaN   \n",
       "\n",
       "                                                   Text  \\\n",
       "0     what is the price for round trip between toron...   \n",
       "1     trying to find a good deal and good timing to ...   \n",
       "2     trying to plan route penn ny to montgomery ny....   \n",
       "3     Hi does the 54 Vermonter train go from Newark,...   \n",
       "4     is there a train service from los angeles ca t...   \n",
       "...                                                 ...   \n",
       "2658                                                NaN   \n",
       "2659                                                NaN   \n",
       "2660                                                NaN   \n",
       "2661                                                NaN   \n",
       "2662                                                NaN   \n",
       "\n",
       "                                       Annotator A Text  \\\n",
       "0     what is the price for round trip between toron...   \n",
       "1     trying to find a good deal and good timing to ...   \n",
       "2     trying to plan route penn ny to montgomery ny....   \n",
       "3     [Hi ]does the 54 Vermonter train go from Newar...   \n",
       "4     is there a train service from los angeles ca t...   \n",
       "...                                                 ...   \n",
       "2658                                                NaN   \n",
       "2659                                                NaN   \n",
       "2660                                                NaN   \n",
       "2661                                                NaN   \n",
       "2662                                                NaN   \n",
       "\n",
       "                                       Annotator B Text  Length  Error  \\\n",
       "0     what is the price for round trip between toron...    61.0    0.0   \n",
       "1     [trying to find a ]good deal and good timing t...    81.0   17.0   \n",
       "2     trying to plan route penn ny to montgomery ny....    63.0   16.0   \n",
       "3     [Hi ]does the 54 Vermonter train go from Newar...    87.0   10.0   \n",
       "4     is there a train service from los angeles ca t...    64.0    0.0   \n",
       "...                                                 ...     ...    ...   \n",
       "2658                                                NaN     NaN    NaN   \n",
       "2659                                                NaN     NaN    NaN   \n",
       "2660                                                NaN     NaN    NaN   \n",
       "2661                                                NaN     NaN    NaN   \n",
       "2662                                                NaN     NaN    NaN   \n",
       "\n",
       "      Alignment Score  Agreement  \n",
       "0            1.000000        1.0  \n",
       "1            0.790123        0.0  \n",
       "2            0.746032        0.0  \n",
       "3            0.885057        1.0  \n",
       "4            1.000000        1.0  \n",
       "...               ...        ...  \n",
       "2658              NaN        NaN  \n",
       "2659              NaN        NaN  \n",
       "2660              NaN        NaN  \n",
       "2661              NaN        NaN  \n",
       "2662              NaN        NaN  \n",
       "\n",
       "[2663 rows x 12 columns]"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.read_csv('data/rsics_dataset/1_2_align.csv')"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "chatbot",
   "language": "python",
   "name": "chatbot"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}