notebooks/02_extract_names_from_news.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "5d9fdc3c-4b89-4019-b85d-a2d825b17f8d",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "\n",
    "import spacy\n",
    "from spacy import displacy\n",
    "\n",
    "from tqdm import tqdm\n",
    "tqdm.pandas()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "451a92bb-fd8a-47f9-a17e-6df29827bd77",
   "metadata": {},
   "outputs": [],
   "source": [
    "nlp = spacy.load(\"en_core_web_lg\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "1c6a57be-1cf4-4b18-b29d-e2eb9a41e10a",
   "metadata": {},
   "outputs": [],
   "source": [
    "articles = !ls -ltr *_articles.csv | awk '{print $9}'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "56c5ff85-17cc-456d-87a6-01ddcc8d3d56",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['cnn_articles.csv',\n",
       " 'abc_articles.csv',\n",
       " 'cbs_articles.csv',\n",
       " 'nbc_articles.csv',\n",
       " 'npr_articles.csv',\n",
       " 'nyt_articles.csv',\n",
       " 'usat_articles.csv',\n",
       " 'wapo_articles.csv',\n",
       " 'politico_articles.csv',\n",
       " 'propub_articles.csv',\n",
       " 'lat_articles.csv']"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "articles"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4f7e7b52-9cce-4193-a81d-eac7bd831134",
   "metadata": {},
   "source": [
    "## Apply NER on all the articles"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "fa46fbde-8863-4245-b8c9-72c5d839414d",
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_names(row):\n",
    "    full_names = set()\n",
    "    try:\n",
    "        for ent in nlp(row).ents:\n",
    "            # check if ent is a full name\n",
    "            if ent.label_ == 'PERSON' and len(ent.text.split()) > 1:\n",
    "                full_names.add(ent.text)\n",
    "    except:\n",
    "        pass\n",
    "    return list(full_names)\n",
    "        \n",
    "\n",
    "def get_names_from_articles(file):\n",
    "    df = pd.read_csv(file)\n",
    "    print(f\"Applying NER to get names for {file}\")\n",
    "    df['names_in_article'] = df['text'].progress_apply(lambda x: get_names(x))\n",
    "    print(f\"writing to csv for {file}\")\n",
    "    df[['title','authors','names_in_article']].to_csv(f\"{file.split('.')[0]}_w_names.csv\",index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "c4576213-3b63-4775-80e6-c06be9df8556",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Applying NER to get names for cnn_articles.csv\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 32958/32958 [1:11:46<00:00,  7.65it/s]  \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "writing to csv for cnn_articles.csv\n",
      "Applying NER to get names for abc_articles.csv\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 62738/62738 [14:59<00:00, 69.75it/s]  \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "writing to csv for abc_articles.csv\n",
      "Applying NER to get names for cbs_articles.csv\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 37907/37907 [1:09:01<00:00,  9.15it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "writing to csv for cbs_articles.csv\n",
      "Applying NER to get names for nbc_articles.csv\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 17968/17968 [33:22<00:00,  8.97it/s]  \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "writing to csv for nbc_articles.csv\n",
      "Applying NER to get names for npr_articles.csv\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 15108/15108 [46:31<00:00,  5.41it/s] \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "writing to csv for npr_articles.csv\n",
      "Applying NER to get names for nyt_articles.csv\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 26738/26738 [33:34<00:00, 13.27it/s]  \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "writing to csv for nyt_articles.csv\n",
      "Applying NER to get names for usat_articles.csv\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 40637/40637 [1:55:25<00:00,  5.87it/s]  \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "writing to csv for usat_articles.csv\n",
      "Applying NER to get names for wapo_articles.csv\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 33764/33764 [1:27:46<00:00,  6.41it/s]  \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "writing to csv for wapo_articles.csv\n",
      "Applying NER to get names for politico_articles.csv\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 2456/2456 [05:24<00:00,  7.57it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "writing to csv for politico_articles.csv\n",
      "Applying NER to get names for propub_articles.csv\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 559/559 [04:23<00:00,  2.12it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "writing to csv for propub_articles.csv\n",
      "Applying NER to get names for lat_articles.csv\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 19510/19510 [59:53<00:00,  5.43it/s]  \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "writing to csv for lat_articles.csv\n"
     ]
    }
   ],
   "source": [
    "for a in articles:\n",
    "    get_names_from_articles(a)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2de652a0-9cf0-48e8-894b-b4f79d7dc84e",
   "metadata": {},
   "source": [
    "## Save all the author names"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "873a4c32-7606-449b-b840-7ba00c48a350",
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_name(text):\n",
    "    name = ''\n",
    "    doc = nlp(text)\n",
    "    for ent in doc.ents:\n",
    "        if ent.label_ == 'PERSON' and len(ent.text.split()) > 1:\n",
    "            name = ent.text\n",
    "    return name"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "id": "15894ca0-6f9e-4b99-adff-0b1ade272609",
   "metadata": {},
   "outputs": [],
   "source": [
    "def save_authors(file):\n",
    "    df = pd.read_csv(file)\n",
    "    print(f\"Processing {file}\")\n",
    "    channel = file.split(\"_\")[0]\n",
    "    merged_df = pd.DataFrame({'full_name': ','.join(df['authors'].dropna()).split(\",\")})\n",
    "    merged_df.drop_duplicates(inplace=True)\n",
    "    merged_df['full_name'] = merged_df['full_name'].str.replace(r'/.*', '', regex=True)\n",
    "    merged_df['full_name'] = merged_df['full_name'].str.replace(r'\\n.*', '', regex=True)\n",
    "    merged_df['full_name'] = merged_df['full_name'].str.replace(\"[^a-zA-Z' -]\", '', regex=True)\n",
    "    merged_df = merged_df.drop(merged_df[merged_df['full_name'].str.len() < 2].index)\n",
    "    merged_df['full_name'] = merged_df['full_name'].progress_apply(lambda x: get_name(x))\n",
    "    print(f\"Writing author names for {channel}\")\n",
    "    merged_df.dropna(inplace=True)\n",
    "    merged_df.drop_duplicates(inplace=True)\n",
    "    merged_df.to_csv(f'{channel}_author_names.csv',index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "08793540-7482-4676-865a-b826d5c57f61",
   "metadata": {},
   "outputs": [],
   "source": [
    "ner_results = !ls -ltr *_w_names.csv | awk '{print $9}'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "2e1ec15e-e3e2-4c36-9e72-e2e641d1ab2a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['cnn_articles_w_names.csv',\n",
       " 'abc_articles_w_names.csv',\n",
       " 'cbs_articles_w_names.csv',\n",
       " 'nbc_articles_w_names.csv',\n",
       " 'npr_articles_w_names.csv',\n",
       " 'nyt_articles_w_names.csv',\n",
       " 'usat_articles_w_names.csv',\n",
       " 'wapo_articles_w_names.csv',\n",
       " 'politico_articles_w_names.csv',\n",
       " 'propub_articles_w_names.csv',\n",
       " 'lat_articles_w_names.csv']"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ner_results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "id": "d290a66a-7ac5-4e2d-ae52-adf6dd734f2a",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Processing cnn_articles_w_names.csv\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 9080/9080 [01:14<00:00, 122.50it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Writing author names for cnn\n",
      "Processing abc_articles_w_names.csv\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 27/27 [00:00<00:00, 115.85it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Writing author names for abc\n",
      "Processing cbs_articles_w_names.csv\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 1285/1285 [00:15<00:00, 83.23it/s] \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Writing author names for cbs\n",
      "Processing nbc_articles_w_names.csv\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 3246/3246 [00:41<00:00, 77.54it/s] \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Writing author names for nbc\n",
      "Processing npr_articles_w_names.csv\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 4375/4375 [00:33<00:00, 128.80it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Writing author names for npr\n",
      "Processing nyt_articles_w_names.csv\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 4637/4637 [00:41<00:00, 110.79it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Writing author names for nyt\n",
      "Processing usat_articles_w_names.csv\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 2533/2533 [00:20<00:00, 123.59it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Writing author names for usat\n",
      "Processing wapo_articles_w_names.csv\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 3748/3748 [00:32<00:00, 116.55it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Writing author names for wapo\n",
      "Processing politico_articles_w_names.csv\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 171/171 [00:01<00:00, 87.44it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Writing author names for politico\n",
      "Processing propub_articles_w_names.csv\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 608/608 [00:04<00:00, 121.85it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Writing author names for propub\n",
      "Processing lat_articles_w_names.csv\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 2992/2992 [00:28<00:00, 106.57it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Writing author names for lat\n"
     ]
    }
   ],
   "source": [
    "for r in ner_results:\n",
    "    save_authors(r)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "id": "45f22741-1127-4c63-a25a-976fb517bbf4",
   "metadata": {},
   "outputs": [],
   "source": [
    "authors_csv = !ls -ltr *_author_names.csv | awk '{print $9}'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "id": "c4a21903-459f-4106-9f28-76dff9ed74e0",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['cnn_author_names.csv',\n",
       " 'abc_author_names.csv',\n",
       " 'cbs_author_names.csv',\n",
       " 'nbc_author_names.csv',\n",
       " 'npr_author_names.csv',\n",
       " 'nyt_author_names.csv',\n",
       " 'usat_author_names.csv',\n",
       " 'wapo_author_names.csv',\n",
       " 'politico_author_names.csv',\n",
       " 'propub_author_names.csv',\n",
       " 'lat_author_names.csv']"
      ]
     },
     "execution_count": 57,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "authors_csv"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "id": "3b516a3a-377d-45f7-bc74-84b5c777035a",
   "metadata": {},
   "outputs": [],
   "source": [
    "def add_channel(file):\n",
    "    df = pd.read_csv(file)\n",
    "    df['channel'] = file.split(\"_\")[0]\n",
    "    return df\n",
    "\n",
    "def merge_all(files, to_file):\n",
    "    dataframes = [add_channel(file) for file in files]\n",
    "    merged_df = pd.concat(dataframes, ignore_index=True)\n",
    "    merged_df.to_csv(to_file, index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "id": "ca53207a-e1f6-490b-b7f8-88a3c6e122f1",
   "metadata": {},
   "outputs": [],
   "source": [
    "merge_all(authors_csv,\"all_authors.csv\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6973146a-daa8-4b4d-90c3-770501d9ee49",
   "metadata": {},
   "source": [
    "## Save all names mentioned in articles by channel"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "id": "c9343f3d-a5e0-495c-8a6e-906d31678aeb",
   "metadata": {},
   "outputs": [],
   "source": [
    "def save_mentioned_names(file):\n",
    "    df = pd.read_csv(file)\n",
    "    print(f\"Processing {file}\")\n",
    "    channel = file.split(\"_\")[0]\n",
    "    merged_df = pd.DataFrame({'full_name': ','.join(df['names_in_article'].dropna()).split(\",\")})\n",
    "    merged_df.drop_duplicates(inplace=True)\n",
    "    merged_df['full_name'] = merged_df['full_name'].str.replace(r'/.*', '', regex=True)\n",
    "    merged_df['full_name'] = merged_df['full_name'].str.replace(r'\\n.*', '', regex=True)\n",
    "    merged_df['full_name'] = merged_df['full_name'].str.replace(r\"^'|'$\", '', regex=True)\n",
    "    merged_df['full_name'] = merged_df['full_name'].str.replace(\"[^a-zA-Z' -]\", '', regex=True)\n",
    "    merged_df['full_name'] = merged_df['full_name'].str.strip()\n",
    "    merged_df['full_name'] = merged_df['full_name'].str.strip(\"'\")\n",
    "    merged_df = merged_df.drop(merged_df[merged_df['full_name'].str.len() < 2].index)\n",
    "    # merged_df['full_name'] = merged_df['full_name'].progress_apply(lambda x: get_name(x))\n",
    "    print(f\"Writing author names for {channel}\")\n",
    "    merged_df.dropna(inplace=True)\n",
    "    merged_df.drop_duplicates(inplace=True)\n",
    "    merged_df.to_csv(f'{channel}_mentioned_names.csv',index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "id": "eb082f68-40ea-4e43-93f3-a77ef7e02407",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Processing cnn_articles_w_names.csv\n",
      "Writing author names for cnn\n",
      "Processing abc_articles_w_names.csv\n",
      "Writing author names for abc\n",
      "Processing cbs_articles_w_names.csv\n",
      "Writing author names for cbs\n",
      "Processing nbc_articles_w_names.csv\n",
      "Writing author names for nbc\n",
      "Processing npr_articles_w_names.csv\n",
      "Writing author names for npr\n",
      "Processing nyt_articles_w_names.csv\n",
      "Writing author names for nyt\n",
      "Processing usat_articles_w_names.csv\n",
      "Writing author names for usat\n",
      "Processing wapo_articles_w_names.csv\n",
      "Writing author names for wapo\n",
      "Processing politico_articles_w_names.csv\n",
      "Writing author names for politico\n",
      "Processing propub_articles_w_names.csv\n",
      "Writing author names for propub\n",
      "Processing lat_articles_w_names.csv\n",
      "Writing author names for lat\n"
     ]
    }
   ],
   "source": [
    "for r in ner_results:\n",
    "    save_mentioned_names(r)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "id": "268538d9-a727-4551-bcfa-57cbac6519a6",
   "metadata": {},
   "outputs": [],
   "source": [
    "mentioned_csv = !ls -ltr *_mentioned_names.csv | awk '{print $9}'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 111,
   "id": "6bf321ba-b49d-43a0-8437-4589b781408e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['cnn_mentioned_names.csv',\n",
       " 'abc_mentioned_names.csv',\n",
       " 'cbs_mentioned_names.csv',\n",
       " 'nbc_mentioned_names.csv',\n",
       " 'npr_mentioned_names.csv',\n",
       " 'nyt_mentioned_names.csv',\n",
       " 'usat_mentioned_names.csv',\n",
       " 'wapo_mentioned_names.csv',\n",
       " 'politico_mentioned_names.csv',\n",
       " 'propub_mentioned_names.csv',\n",
       " 'lat_mentioned_names.csv']"
      ]
     },
     "execution_count": 111,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mentioned_csv"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 112,
   "id": "d76dac30-f89e-438c-b12a-133a21013545",
   "metadata": {},
   "outputs": [],
   "source": [
    "merge_all(mentioned_csv,\"all_mentioned_names.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 113,
   "id": "f8d9c185-9800-409d-8540-c5d531a5a5a0",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "332K\tall_authors.csv\n",
      "9.3M\tall_mentioned_names.csv\n"
     ]
    }
   ],
   "source": [
    "!du -sh all_*.csv"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "32796402-bf71-4f20-a1ef-4a4bb25646c3",
   "metadata": {},
   "source": [
    "## stats by channel"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 114,
   "id": "6ba8eefe-fdb8-4b3e-9cdb-f5c792f6f76c",
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_csv(\"all_mentioned_names.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 130,
   "id": "1ff5b8b0-b261-4b04-9c67-a011a016c5eb",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "df['channel'].value_counts().plot(kind='bar');"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 131,
   "id": "8c5d7fc5-fa28-4d83-b46e-4a09b81b1265",
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_csv(\"all_authors.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 132,
   "id": "12500943-0e1c-4a4f-a7ec-0b6ddc4a6015",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "df['channel'].value_counts().plot(kind='bar');"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6f548262-dd9e-4845-8d39-1eb7d1a8b18e",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": ".venv",
   "language": "python",
   "name": ".venv"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}