analysis/describe_datasets.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Description**: Brief descriptions of the classification tasks used in this experiment.\n",
    "The table is meant to be sent to this [LaTeX table\n",
    "generator](https://www.tablesgenerator.com/latex_tables).\n",
    "\n",
    "**Estimated runtime**: ~10 min b/c it has to locally download datasets, ~5 min if you've\n",
    "already downloaded them"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "from IPython.display import clear_output\n",
    "import pandas as pd\n",
    "from tqdm.auto import tqdm\n",
    "\n",
    "from pretrain_on_test import data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "hf_dataset_names: tuple[str] = tuple(\n",
    "    data.hf_dataset_name_to_classification_dataset_info.keys()\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "25"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(hf_dataset_names)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "def describe(df: pd.DataFrame, random_state: int | None = None):\n",
    "    random_obs = df.sample(n=1, random_state=random_state)\n",
    "    return {\n",
    "        \"lengths\": df[\"text\"].str.len(),\n",
    "        \"num_classes\": len(df[\"label\"].unique()),\n",
    "        \"example\": {\n",
    "            \"text\": random_obs[\"text\"].iloc[0],\n",
    "            \"label\": random_obs[\"label\"].iloc[0],\n",
    "        },\n",
    "    }"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " 96%|█████████▌| 24/25 [04:45<00:27, 27.92s/it]\n"
     ]
    }
   ],
   "source": [
    "descriptions = []\n",
    "_progress_bar = tqdm(enumerate(hf_dataset_names), total=len(hf_dataset_names))\n",
    "for i, dataset in _progress_bar:\n",
    "    clear_output(wait=True)\n",
    "    print(_progress_bar)\n",
    "    classification_dataset_info = data.hf_dataset_name_to_classification_dataset_info[\n",
    "        dataset\n",
    "    ]\n",
    "    descriptions.append(\n",
    "        describe(\n",
    "            data.load_classification_data(classification_dataset_info),\n",
    "            random_state=i,\n",
    "        )\n",
    "    )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "dataset_to_citation = {\n",
    "    \"ag_news\": \"zhang2015character\",\n",
    "    \"SetFit/amazon_counterfactual_en\": \"oneill-etal-2021-wish\",\n",
    "    \"app_reviews\": \"grano2017android\",\n",
    "    \"blog_authorship_corpus\": \"schler2006effects\",\n",
    "    \"christinacdl/clickbait_notclickbait_dataset\": None,\n",
    "    \"climate_fever\": \"diggelmann2020climatefever\",\n",
    "    \"aladar/craigslist_bargains\": \"he-etal-2018-decoupling\",\n",
    "    \"disaster_response_messages\": None,\n",
    "    \"aladar/emo\": \"chatterjee-etal-2019-semeval\",\n",
    "    \"dair-ai/emotion\": \"saravia-etal-2018-carer\",\n",
    "    \"SetFit/enron_spam\": \"metsis2006spam\",\n",
    "    \"financial_phrasebank\": \"Malo2014GoodDO\",\n",
    "    \"classla/FRENK-hate-en\": \"ljubešić2019frenk\",\n",
    "    \"hyperpartisan_news_detection\": \"kiesel-etal-2019-semeval\",\n",
    "    \"limit\": \"manotas-etal-2020-limit\",\n",
    "    \"AmazonScience/massive\": \"fitzgerald-etal-2023-massive\",\n",
    "    \"movie_rationales\": \"deyoung-etal-2020-eraser\",\n",
    "    \"mteb/mtop_domain\": \"muennighoff-etal-2023-mteb\",\n",
    "    \"ccdv/patent-classification\": \"sharma-etal-2019-bigpatent\",\n",
    "    \"rotten_tomatoes\": \"pang-lee-2005-seeing\",\n",
    "    \"silicone\": \"chapuis-etal-2020-hierarchical\",\n",
    "    \"trec\": \"wang-etal-2007-jeopardy\",\n",
    "    \"tweets_hate_speech_detection\": \"sharma2019\",\n",
    "    \"yahoo_answers_topics\": \"zhang2018\",\n",
    "    \"yelp_review_full\": \"zhang2015character\",\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "assert set(dataset_to_citation.keys()) == set(hf_dataset_names)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "max_text_length_displayed = 30\n",
    "records = []\n",
    "for dataset, description in zip(hf_dataset_names, descriptions):\n",
    "    dataset_link = f\"https://huggingface.co/datasets/{dataset}\"\n",
    "    dataset_tex = \"\\\\texttt{\" + dataset.replace(\"_\", \"\\_\") + \"}\"\n",
    "    name_and_link = f\"\\\\href{{{dataset_link}}}{{{dataset_tex}}}\"\n",
    "    citation = dataset_to_citation[dataset]\n",
    "    if citation is None:\n",
    "        citation_tex = \"\"\n",
    "    else:\n",
    "        citation_tex = f\"\\\\citet{{{citation}}}\"\n",
    "    records.append(\n",
    "        {\n",
    "            \"Hugging Face dataset\": name_and_link,\n",
    "            \"Author(s)\": citation_tex,\n",
    "            \"Number of classes\": description[\"num_classes\"],\n",
    "            \"Text length (25, 75) percentiles\": tuple(\n",
    "                description[\"lengths\"].describe()[[\"25%\", \"75%\"]].astype(int).values\n",
    "            ),\n",
    "            \"Example text (truncated)\": description[\"example\"][\"text\"][:max_text_length_displayed],\n",
    "        }\n",
    "    )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.DataFrame(records)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "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>Hugging Face dataset</th>\n",
       "      <th>Author(s)</th>\n",
       "      <th>Number of classes</th>\n",
       "      <th>Text length (25, 75) percentiles</th>\n",
       "      <th>Example text (truncated)</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>\\href{https://huggingface.co/datasets/ag_news}...</td>\n",
       "      <td>\\citet{zhang2015character}</td>\n",
       "      <td>4</td>\n",
       "      <td>(196, 266)</td>\n",
       "      <td>First class to the moon London</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>\\href{https://huggingface.co/datasets/SetFit/a...</td>\n",
       "      <td>\\citet{oneill-etal-2021-wish}</td>\n",
       "      <td>2</td>\n",
       "      <td>(60, 125)</td>\n",
       "      <td>I don't know why X-Mini doesn'</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>\\href{https://huggingface.co/datasets/app_revi...</td>\n",
       "      <td>\\citet{grano2017android}</td>\n",
       "      <td>5</td>\n",
       "      <td>(10, 77)</td>\n",
       "      <td>watshpp</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>\\href{https://huggingface.co/datasets/blog_aut...</td>\n",
       "      <td>\\citet{schler2006effects}</td>\n",
       "      <td>2</td>\n",
       "      <td>(92, 556)</td>\n",
       "      <td>\"Why did you date him again? W</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>\\href{https://huggingface.co/datasets/christin...</td>\n",
       "      <td></td>\n",
       "      <td>2</td>\n",
       "      <td>(46, 69)</td>\n",
       "      <td>South China Sea: Beijing says</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>\\href{https://huggingface.co/datasets/climate_...</td>\n",
       "      <td>\\citet{diggelmann2020climatefever}</td>\n",
       "      <td>4</td>\n",
       "      <td>(80, 156)</td>\n",
       "      <td>'Our harmless emissions of tri</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>\\href{https://huggingface.co/datasets/aladar/c...</td>\n",
       "      <td>\\citet{he-etal-2018-decoupling}</td>\n",
       "      <td>6</td>\n",
       "      <td>(346, 713)</td>\n",
       "      <td>Buyer: I am very interested in</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>\\href{https://huggingface.co/datasets/disaster...</td>\n",
       "      <td></td>\n",
       "      <td>3</td>\n",
       "      <td>(74, 178)</td>\n",
       "      <td>A week later, another thunders</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>\\href{https://huggingface.co/datasets/emo}{\\te...</td>\n",
       "      <td>\\citet{chatterjee-etal-2019-semeval}</td>\n",
       "      <td>4</td>\n",
       "      <td>(44, 83)</td>\n",
       "      <td>get lost lost with you sure go</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>\\href{https://huggingface.co/datasets/dair-ai/...</td>\n",
       "      <td>\\citet{saravia-etal-2018-carer}</td>\n",
       "      <td>6</td>\n",
       "      <td>(53, 129)</td>\n",
       "      <td>i just remember feeling really</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>\\href{https://huggingface.co/datasets/SetFit/e...</td>\n",
       "      <td>\\citet{metsis2006spam}</td>\n",
       "      <td>2</td>\n",
       "      <td>(342, 1553)</td>\n",
       "      <td>tenaska iv feb 2001 we do not</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>\\href{https://huggingface.co/datasets/financia...</td>\n",
       "      <td>\\citet{Malo2014GoodDO}</td>\n",
       "      <td>3</td>\n",
       "      <td>(79, 157)</td>\n",
       "      <td>` Sanoma is a buyer not a targ</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>\\href{https://huggingface.co/datasets/classla/...</td>\n",
       "      <td>\\citet{ljubešić2019frenk}</td>\n",
       "      <td>2</td>\n",
       "      <td>(34, 160)</td>\n",
       "      <td>their struck for money as it i</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>\\href{https://huggingface.co/datasets/hyperpar...</td>\n",
       "      <td>\\citet{kiesel-etal-2019-semeval}</td>\n",
       "      <td>2</td>\n",
       "      <td>(39, 63)</td>\n",
       "      <td>Europe ETFs: Opportunity Await</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>\\href{https://huggingface.co/datasets/limit}{\\...</td>\n",
       "      <td>\\citet{manotas-etal-2020-limit}</td>\n",
       "      <td>2</td>\n",
       "      <td>(53, 123)</td>\n",
       "      <td>The man points to notes writ</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>\\href{https://huggingface.co/datasets/AmazonSc...</td>\n",
       "      <td>\\citet{fitzgerald-etal-2023-massive}</td>\n",
       "      <td>18</td>\n",
       "      <td>(24, 44)</td>\n",
       "      <td>turn off all alarms</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>\\href{https://huggingface.co/datasets/movie_ra...</td>\n",
       "      <td>\\citet{deyoung-etal-2020-eraser}</td>\n",
       "      <td>2</td>\n",
       "      <td>(2721, 4659)</td>\n",
       "      <td>there was a lot riding on this</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>\\href{https://huggingface.co/datasets/mteb/mto...</td>\n",
       "      <td>\\citet{muennighoff-etal-2023-mteb}</td>\n",
       "      <td>11</td>\n",
       "      <td>(26, 44)</td>\n",
       "      <td>play that tool song again</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>\\href{https://huggingface.co/datasets/ccdv/pat...</td>\n",
       "      <td>\\citet{sharma-etal-2019-bigpatent}</td>\n",
       "      <td>9</td>\n",
       "      <td>(441, 775)</td>\n",
       "      <td>this invention relates to a me</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>\\href{https://huggingface.co/datasets/rotten_t...</td>\n",
       "      <td>\\citet{pang-lee-2005-seeing}</td>\n",
       "      <td>2</td>\n",
       "      <td>(76, 149)</td>\n",
       "      <td>there are now two signs that m</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>\\href{https://huggingface.co/datasets/silicone...</td>\n",
       "      <td>\\citet{chapuis-etal-2020-hierarchical}</td>\n",
       "      <td>4</td>\n",
       "      <td>(29, 75)</td>\n",
       "      <td>... i know i'll be killed if i</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>\\href{https://huggingface.co/datasets/trec}{\\t...</td>\n",
       "      <td>\\citet{wang-etal-2007-jeopardy}</td>\n",
       "      <td>6</td>\n",
       "      <td>(36, 61)</td>\n",
       "      <td>What country did King Wencesla</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>\\href{https://huggingface.co/datasets/tweets_h...</td>\n",
       "      <td>\\citet{sharma2019}</td>\n",
       "      <td>2</td>\n",
       "      <td>(62, 107)</td>\n",
       "      <td>how did my   memories from rom</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>\\href{https://huggingface.co/datasets/yahoo_an...</td>\n",
       "      <td>\\citet{zhang2018}</td>\n",
       "      <td>10</td>\n",
       "      <td>(58, 213)</td>\n",
       "      <td>How do you see the good in eac</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>\\href{https://huggingface.co/datasets/yelp_rev...</td>\n",
       "      <td>\\citet{zhang2015character}</td>\n",
       "      <td>5</td>\n",
       "      <td>(287, 957)</td>\n",
       "      <td>Fun place for fresh produce or</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                 Hugging Face dataset  \\\n",
       "0   \\href{https://huggingface.co/datasets/ag_news}...   \n",
       "1   \\href{https://huggingface.co/datasets/SetFit/a...   \n",
       "2   \\href{https://huggingface.co/datasets/app_revi...   \n",
       "3   \\href{https://huggingface.co/datasets/blog_aut...   \n",
       "4   \\href{https://huggingface.co/datasets/christin...   \n",
       "5   \\href{https://huggingface.co/datasets/climate_...   \n",
       "6   \\href{https://huggingface.co/datasets/aladar/c...   \n",
       "7   \\href{https://huggingface.co/datasets/disaster...   \n",
       "8   \\href{https://huggingface.co/datasets/emo}{\\te...   \n",
       "9   \\href{https://huggingface.co/datasets/dair-ai/...   \n",
       "10  \\href{https://huggingface.co/datasets/SetFit/e...   \n",
       "11  \\href{https://huggingface.co/datasets/financia...   \n",
       "12  \\href{https://huggingface.co/datasets/classla/...   \n",
       "13  \\href{https://huggingface.co/datasets/hyperpar...   \n",
       "14  \\href{https://huggingface.co/datasets/limit}{\\...   \n",
       "15  \\href{https://huggingface.co/datasets/AmazonSc...   \n",
       "16  \\href{https://huggingface.co/datasets/movie_ra...   \n",
       "17  \\href{https://huggingface.co/datasets/mteb/mto...   \n",
       "18  \\href{https://huggingface.co/datasets/ccdv/pat...   \n",
       "19  \\href{https://huggingface.co/datasets/rotten_t...   \n",
       "20  \\href{https://huggingface.co/datasets/silicone...   \n",
       "21  \\href{https://huggingface.co/datasets/trec}{\\t...   \n",
       "22  \\href{https://huggingface.co/datasets/tweets_h...   \n",
       "23  \\href{https://huggingface.co/datasets/yahoo_an...   \n",
       "24  \\href{https://huggingface.co/datasets/yelp_rev...   \n",
       "\n",
       "                                 Author(s)  Number of classes  \\\n",
       "0               \\citet{zhang2015character}                  4   \n",
       "1            \\citet{oneill-etal-2021-wish}                  2   \n",
       "2                 \\citet{grano2017android}                  5   \n",
       "3                \\citet{schler2006effects}                  2   \n",
       "4                                                           2   \n",
       "5       \\citet{diggelmann2020climatefever}                  4   \n",
       "6          \\citet{he-etal-2018-decoupling}                  6   \n",
       "7                                                           3   \n",
       "8     \\citet{chatterjee-etal-2019-semeval}                  4   \n",
       "9          \\citet{saravia-etal-2018-carer}                  6   \n",
       "10                  \\citet{metsis2006spam}                  2   \n",
       "11                  \\citet{Malo2014GoodDO}                  3   \n",
       "12               \\citet{ljubešić2019frenk}                  2   \n",
       "13        \\citet{kiesel-etal-2019-semeval}                  2   \n",
       "14         \\citet{manotas-etal-2020-limit}                  2   \n",
       "15    \\citet{fitzgerald-etal-2023-massive}                 18   \n",
       "16        \\citet{deyoung-etal-2020-eraser}                  2   \n",
       "17      \\citet{muennighoff-etal-2023-mteb}                 11   \n",
       "18      \\citet{sharma-etal-2019-bigpatent}                  9   \n",
       "19            \\citet{pang-lee-2005-seeing}                  2   \n",
       "20  \\citet{chapuis-etal-2020-hierarchical}                  4   \n",
       "21         \\citet{wang-etal-2007-jeopardy}                  6   \n",
       "22                      \\citet{sharma2019}                  2   \n",
       "23                       \\citet{zhang2018}                 10   \n",
       "24              \\citet{zhang2015character}                  5   \n",
       "\n",
       "   Text length (25, 75) percentiles        Example text (truncated)  \n",
       "0                        (196, 266)  First class to the moon London  \n",
       "1                         (60, 125)  I don't know why X-Mini doesn'  \n",
       "2                          (10, 77)                         watshpp  \n",
       "3                         (92, 556)  \"Why did you date him again? W  \n",
       "4                          (46, 69)  South China Sea: Beijing says   \n",
       "5                         (80, 156)  'Our harmless emissions of tri  \n",
       "6                        (346, 713)  Buyer: I am very interested in  \n",
       "7                         (74, 178)  A week later, another thunders  \n",
       "8                          (44, 83)  get lost lost with you sure go  \n",
       "9                         (53, 129)  i just remember feeling really  \n",
       "10                      (342, 1553)  tenaska iv feb 2001 we do not   \n",
       "11                        (79, 157)  ` Sanoma is a buyer not a targ  \n",
       "12                        (34, 160)  their struck for money as it i  \n",
       "13                         (39, 63)  Europe ETFs: Opportunity Await  \n",
       "14                        (53, 123)    The man points to notes writ  \n",
       "15                         (24, 44)             turn off all alarms  \n",
       "16                     (2721, 4659)  there was a lot riding on this  \n",
       "17                         (26, 44)       play that tool song again  \n",
       "18                       (441, 775)  this invention relates to a me  \n",
       "19                        (76, 149)  there are now two signs that m  \n",
       "20                         (29, 75)  ... i know i'll be killed if i  \n",
       "21                         (36, 61)  What country did King Wencesla  \n",
       "22                        (62, 107)  how did my   memories from rom  \n",
       "23                        (58, 213)  How do you see the good in eac  \n",
       "24                       (287, 957)  Fun place for fresh produce or  "
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\\href{https://huggingface.co/datasets/ag_news}{\\texttt{ag\\_news}}\n"
     ]
    }
   ],
   "source": [
    "print(df[\"Hugging Face dataset\"].iloc[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\\citet{zhang2015character}\n"
     ]
    }
   ],
   "source": [
    "print(df[\"Author(s)\"].iloc[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(2, 18)"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df[\"Number of classes\"].min(), df[\"Number of classes\"].max()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "# df.to_csv(\"dataset_descriptions.csv\", index=False)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "pretrain",
   "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": "3.11.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}