w2_text_extract_crcp (1).ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# PDF Extraction Demo"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# (1) Extract each page of the PDF CRCP flyer and convert it to a separage image file"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "import fitz"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "filename = \"CRCP_brochure.pdf\"\n",
    "doc = fitz.open(filename)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Loop through each page in the doc and extract pixel map (sets of pixels defining its color), save as a png file\n",
    "for each in range(len(doc)):\n",
    "    doc.load_page(each).get_pixmap().save(f\"crcp_page{each+1}.png\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# (2) Extract the text from the first page of the CRCP flyer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "from transformers import LayoutLMv3FeatureExtractor \n",
    "from PIL import Image\n",
    "import pandas as pd\n",
    "import glob"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Open the image of the first page (page 1)\n",
    "image = Image.open(\"crcp_page1.png\").convert(\"RGB\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Instantiate a transformers deep learning model \n",
    "pytorch_text_extractor = LayoutLMv3FeatureExtractor()\n",
    "\n",
    "# Use the model to extract text from the image of the first page of the flyer\n",
    "encoding = pytorch_text_extractor(image, return_tensors=\"pt\") # returns PyTorch tensors"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[['FINRA',\n",
       "  'Institute',\n",
       "  'at',\n",
       "  'Georgetown',\n",
       "  '=',\n",
       "  'The',\n",
       "  'Securities',\n",
       "  \"Industry's\",\n",
       "  'Leading',\n",
       "  'Executive',\n",
       "  'Education',\n",
       "  'Program',\n",
       "  'for',\n",
       "  'Compliance',\n",
       "  'Professionals',\n",
       "  'GEORGETOWN',\n",
       "  'UNIVERSITY',\n",
       "  'McDonough',\n",
       "  'ScHoot',\n",
       "  'of',\n",
       "  'BusINEss']]"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "encoding[\"words\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Import libraries to make a word cloud\n",
    "from wordcloud import WordCloud, STOPWORDS\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "wc = WordCloud().generate(\" \".join(encoding[\"words\"][0]))\n",
    "plt.imshow(wc, interpolation='bilinear')\n",
    "plt.axis(\"off\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# (3) Extract the text from all pages of the CRCP flyer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "import glob\n",
    "import os"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Create list of documents in working directory (current folder)\n",
    "# that start with \"crcp\", order it by modification time\n",
    "doc_images = sorted(glob.glob(\"crcp*\"), key=os.path.getmtime)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['crcp_page1.png',\n",
       " 'crcp_page2.png',\n",
       " 'crcp_page3.png',\n",
       " 'crcp_page4.png',\n",
       " 'crcp_page5.png']"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "doc_images"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "## Loop through each page and extract the text on the page\n",
    "\n",
    "encodings = []\n",
    "\n",
    "for each in doc_images:\n",
    "    image = Image.open(each).convert(\"RGB\")\n",
    "    pytorch_text_extractor = LayoutLMv3FeatureExtractor()\n",
    "    encodings.append(pytorch_text_extractor(image, return_tensors=\"pt\")) # returns PyTorch tensors"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[['Program',\n",
       "  'Elements',\n",
       "  'and',\n",
       "  'Requirements',\n",
       "  '‘Week',\n",
       "  'I:',\n",
       "  'Foundation',\n",
       "  '|',\n",
       "  'July',\n",
       "  '17-22,',\n",
       "  '2022',\n",
       "  'Participants',\n",
       "  'gather',\n",
       "  'on',\n",
       "  'Georgetown',\n",
       "  'University’s',\n",
       "  'campus',\n",
       "  'at',\n",
       "  'the',\n",
       "  'McDonough',\n",
       "  'School',\n",
       "  '(of',\n",
       "  'Business',\n",
       "  'for',\n",
       "  'intensive',\n",
       "  'instruction,',\n",
       "  'discussion',\n",
       "  'and',\n",
       "  'case',\n",
       "  'studies,',\n",
       "  'gaining',\n",
       "  'a',\n",
       "  'strong',\n",
       "  'Understanding',\n",
       "  'of',\n",
       "  'the',\n",
       "  'foundation',\n",
       "  'and',\n",
       "  'application',\n",
       "  'of',\n",
       "  'the',\n",
       "  'financial',\n",
       "  'services',\n",
       "  'industry',\n",
       "  'rules',\n",
       "  'and',\n",
       "  'financial',\n",
       "  'regulations.',\n",
       "  'Topics',\n",
       "  'include:',\n",
       "  '>',\n",
       "  'Securities',\n",
       "  'law',\n",
       "  'and',\n",
       "  'regulatory',\n",
       "  'structure»',\n",
       "  'Books',\n",
       "  'and',\n",
       "  'records',\n",
       "  '>',\n",
       "  'Supervisory',\n",
       "  'practices',\n",
       "  'and',\n",
       "  'internal',\n",
       "  'controls»',\n",
       "  'Communications',\n",
       "  'withthe',\n",
       "  'public',\n",
       "  '‘Week',\n",
       "  'Il:',\n",
       "  'Capstone',\n",
       "  '|',\n",
       "  'November',\n",
       "  '13',\n",
       "  '~',\n",
       "  '18,',\n",
       "  '2022',\n",
       "  'Participants',\n",
       "  'reconvene',\n",
       "  'at',\n",
       "  'Georgetown',\n",
       "  'University’s',\n",
       "  'McDonough',\n",
       "  'School',\n",
       "  'of',\n",
       "  'Business',\n",
       "  'for',\n",
       "  'in-depth',\n",
       "  'coursework',\n",
       "  'that',\n",
       "  'explores',\n",
       "  'a',\n",
       "  'series',\n",
       "  'of',\n",
       "  'advanced-level',\n",
       "  'regulatory',\n",
       "  'and',\n",
       "  'compliance',\n",
       "  '‘topics',\n",
       "  'that',\n",
       "  'help',\n",
       "  'participants',\n",
       "  'build',\n",
       "  'on',\n",
       "  'previous',\n",
       "  'coursework',\n",
       "  'covered',\n",
       "  'during',\n",
       "  'Week',\n",
       "  'lof',\n",
       "  'the',\n",
       "  'ERCP',\n",
       "  'program.',\n",
       "  'Topics',\n",
       "  'include:',\n",
       "  '>',\n",
       "  'Cybersecurity',\n",
       "  '>',\n",
       "  'Regulation',\n",
       "  'Best',\n",
       "  'interest',\n",
       "  '>',\n",
       "  'Antiemoney',\n",
       "  'laundering»',\n",
       "  'Fintech',\n",
       "  'Although',\n",
       "  'coursework',\n",
       "  'must',\n",
       "  'be',\n",
       "  'completed',\n",
       "  'within',\n",
       "  'two',\n",
       "  'consecutive',\n",
       "  'calendar',\n",
       "  'years,',\n",
       "  '‘completion',\n",
       "  'in',\n",
       "  'the',\n",
       "  'same',\n",
       "  'calendar',\n",
       "  'year',\n",
       "  'is',\n",
       "  'recommended.',\n",
       "  'Pre-Course',\n",
       "  'Work',\n",
       "  'Candidates',\n",
       "  'are',\n",
       "  'assigned',\n",
       "  'materials',\n",
       "  'to',\n",
       "  'review',\n",
       "  'in',\n",
       "  'advance',\n",
       "  'ofthe',\n",
       "  'coursework—',\n",
       "  'including',\n",
       "  'rules,',\n",
       "  'regulations,',\n",
       "  'compliance',\n",
       "  'and',\n",
       "  'regulatory',\n",
       "  'guidance,',\n",
       "  'and',\n",
       "  'other',\n",
       "  'relevant',\n",
       "  'readings.',\n",
       "  'Assessments',\n",
       "  '‘On',\n",
       "  'the',\n",
       "  'final',\n",
       "  'day',\n",
       "  'ofthe',\n",
       "  'Week',\n",
       "  '|',\n",
       "  'and',\n",
       "  'Week',\n",
       "  'Il',\n",
       "  'courses,',\n",
       "  'participants',\n",
       "  'must',\n",
       "  'complete',\n",
       "  'a',\n",
       "  'written',\n",
       "  'assessment.',\n",
       "  'Participants',\n",
       "  'who',\n",
       "  'pass',\n",
       "  'both',\n",
       "  'assessments',\n",
       "  'are',\n",
       "  'awarded',\n",
       "  'the',\n",
       "  'CRCP',\n",
       "  'designation,',\n",
       "  'con',\n",
       "  '1g',\n",
       "  'Education',\n",
       "  'Program',\n",
       "  'graduates',\n",
       "  'ate',\n",
       "  'required',\n",
       "  'to',\n",
       "  'complete',\n",
       "  '12',\n",
       "  'hours',\n",
       "  'of',\n",
       "  'continuing',\n",
       "  'education',\n",
       "  'every',\n",
       "  'three',\n",
       "  'years',\n",
       "  'as',\n",
       "  'of',\n",
       "  'their',\n",
       "  'certificate',\n",
       "  'anniversary',\n",
       "  'award',\n",
       "  'date.',\n",
       "  'The',\n",
       "  'continuing',\n",
       "  'education',\n",
       "  'requirement',\n",
       "  'helps',\n",
       "  'ensure',\n",
       "  'that',\n",
       "  'graduates',\n",
       "  'remain',\n",
       "  'current',\n",
       "  'on',\n",
       "  'compliance',\n",
       "  'topics,',\n",
       "  'rules',\n",
       "  'and',\n",
       "  'regulations,',\n",
       "  'and',\n",
       "  'emerging',\n",
       "  'issues',\n",
       "  'in',\n",
       "  'the',\n",
       "  'financial',\n",
       "  'services',\n",
       "  'industry.']]"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# SHOW: 0: first page, 1: second page, etc...\n",
    "encodings[2][\"words\"]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# (4) Process extracted text to analyze"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Combine text from each page into a single string\n",
    "all_text = [\" \".join(each[\"words\"][0]) for each in encodings]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Create a dataframe to store all_text, each row is a page in the PDF\n",
    "finra_df = pd.DataFrame({\"page\":range(1,6),\n",
    "                         \"raw_text\":all_text})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "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>page</th>\n",
       "      <th>raw_text</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>FINRA Institute at Georgetown = The Securities...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>Become a Certified Regulatory and Compliance P...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3</td>\n",
       "      <td>Program Elements and Requirements ‘Week I: Fou...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4</td>\n",
       "      <td>Registration &amp; Fees ‘To register, complete the...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5</td>\n",
       "      <td>About FINRA and Georgetown University’s McDono...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   page                                           raw_text\n",
       "0     1  FINRA Institute at Georgetown = The Securities...\n",
       "1     2  Become a Certified Regulatory and Compliance P...\n",
       "2     3  Program Elements and Requirements ‘Week I: Fou...\n",
       "3     4  Registration & Fees ‘To register, complete the...\n",
       "4     5  About FINRA and Georgetown University’s McDono..."
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "finra_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 7 Clean text\n",
    "def process_text(t):\n",
    "    import re\n",
    "    \n",
    "    t = t.lower()\n",
    "    t = re.sub(r\"http\\S+\", \"\", t)      # remove links (anything that doesn't have a space after http)\n",
    "    t = re.sub(r\"www.\\S+\", \"\", t)      # remove links\n",
    "    t = re.sub(\"[^a-z]\", \" \", t)\n",
    "    return t"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "finra_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Apply process_text function to raw_text column\n",
    "finra_df[\"clean_text\"] = finra_df[\"raw_text\"].apply(process_text)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "finra_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Import libraries to make a word cloud\n",
    "from wordcloud import WordCloud, STOPWORDS\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "wc = WordCloud(stopwords=STOPWORDS).generate(\" \".join(finra_df[\"clean_text\"]))\n",
    "plt.imshow(wc, interpolation='bilinear')\n",
    "plt.axis(\"off\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.feature_extraction.text import CountVectorizer\n",
    "import altair as alt\n",
    "\n",
    "# 1 Instantiates a vectorizer to vectorize the text (extract features) based on ngram parameters we specify\n",
    "v = CountVectorizer(stop_words=\"english\",\n",
    "                    ngram_range=(1, 1))\n",
    "    \n",
    "# 2 Vectorizes the text, creates a dataframe with terms and frequencies\n",
    "dtm = v.fit_transform(finra_df[\"clean_text\"])\n",
    "    \n",
    "dtm_df = pd.DataFrame(dtm.toarray(),\n",
    "                     columns=v.get_feature_names_out())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "dtm_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Transposes the dataframe, make index a column, and renames columns so it is two columns: text, freq\n",
    "dtm_df = dtm_df.T"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "dtm_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Make term a column not the index\n",
    "dtm_df.reset_index(inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "dtm_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Rename the columns\n",
    "dtm_df.columns = [\"term\", \"page1\", \"page2\", \"page3\", \"page4\", \"page5\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "dtm_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "dtm_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Plot words used on each page\n",
    "alt.Chart(dtm_df[dtm_df[\"page3\"] > 0]).mark_bar().encode(alt.Y(\"term\", sort=\"-x\"),\n",
    "                                      x = \"page3\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# (5) Define a function to adjust the settings/phrases"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "def crcp_words(df, page, phrase_min, phrase_max):\n",
    "    \n",
    "    from sklearn.feature_extraction.text import CountVectorizer\n",
    "    import altair as alt\n",
    "\n",
    "    # 1 Instantiates a vectorizer to vectorize the text (extract features) based on ngram parameters we specify\n",
    "    v = CountVectorizer(stop_words=\"english\",\n",
    "                    ngram_range=(phrase_min, phrase_max))\n",
    "    \n",
    "    # 2 Vectorizes the text, creates a dataframe with terms and frequencies\n",
    "    dtm = v.fit_transform(df[\"clean_text\"])\n",
    "    \n",
    "    dtm_df = pd.DataFrame(dtm.toarray(),\n",
    "                     columns=v.get_feature_names_out())\n",
    "    \n",
    "    dtm_df = dtm_df.T\n",
    "    dtm_df.reset_index(inplace=True)\n",
    "    dtm_df.columns = [\"term\", \"page1\", \"page2\", \"page3\", \"page4\", \"page5\"]\n",
    "    \n",
    "    plot = alt.Chart(dtm_df[dtm_df[f\"{page}\"] > 1]).mark_bar().encode(alt.Y(\"term\", sort=\"-x\"),\n",
    "                                      x = f\"{page}\")\n",
    "    \n",
    "    return plot"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "dtm_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Use the function\n",
    "crcp_words(finra_df, page=\"page4\", phrase_min = 3, phrase_max = 3)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    " ## Bonus: Extract text from screenshots"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "##### Took screenshot of SEC site"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Open the image \n",
    "image = Image.open(\"sec_shot.png\").convert(\"RGB\")\n",
    "\n",
    "# Instantiate a transformers deep learning model \n",
    "pytorch_text_extractor = LayoutLMv3FeatureExtractor()\n",
    "\n",
    "# Use the model to extract text from the image of the first page of the site\n",
    "encoding = pytorch_text_extractor(image, return_tensors=\"pt\") # returns PyTorch tensors"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[['EDGAR',\n",
       "  '-',\n",
       "  'Search',\n",
       "  'and',\n",
       "  'Access',\n",
       "  'Latest',\n",
       "  'Filings',\n",
       "  'Company',\n",
       "  'Filings',\n",
       "  'Mutual',\n",
       "  'Funds',\n",
       "  'Variable',\n",
       "  'Insurance',\n",
       "  'Products',\n",
       "  'Daily',\n",
       "  'Filings',\n",
       "  'by',\n",
       "  'Type',\n",
       "  'EDGAR',\n",
       "  'Full',\n",
       "  'Text',\n",
       "  'Search',\n",
       "  'CIK',\n",
       "  'Lookup',\n",
       "  'Confidential',\n",
       "  'Treatment',\n",
       "  'Orders',\n",
       "  'Effectiveness',\n",
       "  'Notices',\n",
       "  'Aknt',\n",
       "  'CReAD',\n",
       "  'EDGAR—Search',\n",
       "  'and',\n",
       "  'Access',\n",
       "  'You',\n",
       "  'can',\n",
       "  'search',\n",
       "  'information',\n",
       "  'collected',\n",
       "  'by',\n",
       "  'the',\n",
       "  'SEC',\n",
       "  'using',\n",
       "  'a',\n",
       "  'variety',\n",
       "  'of',\n",
       "  'search',\n",
       "  'tools.',\n",
       "  'EDGAR',\n",
       "  'full',\n",
       "  'text',\n",
       "  'search',\n",
       "  '¢',\n",
       "  'New',\n",
       "  'versatile',\n",
       "  'tool',\n",
       "  'lets',\n",
       "  'you',\n",
       "  'search',\n",
       "  'for',\n",
       "  'keywords',\n",
       "  'and',\n",
       "  'phrases',\n",
       "  'in',\n",
       "  'over',\n",
       "  '20',\n",
       "  'years',\n",
       "  'of',\n",
       "  'EDGAR',\n",
       "  'filings,',\n",
       "  'and',\n",
       "  'filter',\n",
       "  'by',\n",
       "  'date,',\n",
       "  'company,',\n",
       "  'person,',\n",
       "  'filing',\n",
       "  'category,',\n",
       "  'or',\n",
       "  'location.',\n",
       "  'Boolean',\n",
       "  'and',\n",
       "  'advanced',\n",
       "  'searching,',\n",
       "  'including',\n",
       "  'addresses',\n",
       "  'Search',\n",
       "  'by:',\n",
       "  '¢',\n",
       "  'Company',\n",
       "  'or',\n",
       "  'fund',\n",
       "  'name,',\n",
       "  'ticker',\n",
       "  'symbol,',\n",
       "  'central',\n",
       "  'index',\n",
       "  'key',\n",
       "  '(CIK),',\n",
       "  'file',\n",
       "  'number,',\n",
       "  'state,',\n",
       "  'country,',\n",
       "  'or',\n",
       "  'standard',\n",
       "  'industrial',\n",
       "  'classification',\n",
       "  '(SIC)',\n",
       "  '2006',\n",
       "  '¢',\n",
       "  'Variable',\n",
       "  'insurance',\n",
       "  'products',\n",
       "  'by',\n",
       "  'name',\n",
       "  'of',\n",
       "  'insurance',\n",
       "  'company,',\n",
       "  'underlying',\n",
       "  'mutual',\n",
       "  'fund,',\n",
       "  'or',\n",
       "  'contract,',\n",
       "  'since',\n",
       "  'February',\n",
       "  '2006',\n",
       "  'Search',\n",
       "  'for:',\n",
       "  '¢',\n",
       "  \"Acompany's\",\n",
       "  'central',\n",
       "  'index',\n",
       "  'key',\n",
       "  '(CIK)',\n",
       "  'number',\n",
       "  '¢',\n",
       "  'Latest',\n",
       "  'filings',\n",
       "  'Oao08',\n",
       "  'Resources',\n",
       "  'Researching',\n",
       "  'Public',\n",
       "  'Companies',\n",
       "  'Through',\n",
       "  'EDGAR:',\n",
       "  'A',\n",
       "  'Guide',\n",
       "  'for',\n",
       "  'Investors',\n",
       "  'Tips',\n",
       "  'for',\n",
       "  'using',\n",
       "  'EDGAR',\n",
       "  'and',\n",
       "  'answers',\n",
       "  'to',\n",
       "  'frequently',\n",
       "  'asked',\n",
       "  'questions',\n",
       "  'about',\n",
       "  'researching',\n",
       "  'public',\n",
       "  'companies',\n",
       "  'New',\n",
       "  'data',\n",
       "  'APIs',\n",
       "  'now',\n",
       "  'available',\n",
       "  'Accessing',\n",
       "  'EDGAR',\n",
       "  'Data',\n",
       "  'Resources',\n",
       "  'for',\n",
       "  'developers',\n",
       "  'EDGAR',\n",
       "  'Indexes—Daily,',\n",
       "  'Full',\n",
       "  'and',\n",
       "  'Quarterly',\n",
       "  'Paper',\n",
       "  'Forms',\n",
       "  '144',\n",
       "  'submitted',\n",
       "  'via',\n",
       "  'email.',\n",
       "  'For',\n",
       "  'members',\n",
       "  'of',\n",
       "  'the',\n",
       "  'public',\n",
       "  'who',\n",
       "  'wish',\n",
       "  'to',\n",
       "  'review',\n",
       "  'Forms',\n",
       "  '144',\n",
       "  'filed',\n",
       "  'in',\n",
       "  'paper,',\n",
       "  'the',\n",
       "  'Public',\n",
       "  'Reference',\n",
       "  'Room',\n",
       "  'is',\n",
       "  'open',\n",
       "  'on',\n",
       "  'Wednesdays',\n",
       "  'from',\n",
       "  '10',\n",
       "  'a.m.',\n",
       "  '—',\n",
       "  '3:30',\n",
       "  'p.m.',\n",
       "  'ET.',\n",
       "  'Certified',\n",
       "  'EDGAR',\n",
       "  'Records—order',\n",
       "  'and',\n",
       "  'requect',\n",
       "  'electronic',\n",
       "  'certified']]"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "encoding[\"words\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "pd.set_option(\"display.max_colwidth\", None)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}