{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "27d3ad11",
   "metadata": {
    "vscode": {
     "languageId": "plaintext"
    }
   },
   "outputs": [],
   "source": [
    "# Install dependencies\n",
    "!pip install datasets pandas\n",
    "\n",
    "from datasets import load_dataset\n",
    "import pandas as pd\n",
    "\n",
    "# Load NOTEEVENTS directly from Hugging Face (no manual download)\n",
    "notes = load_dataset(\n",
    "    \"ntphuc149/MIMIC-III-Clinical-Database\",\n",
    "    \"NOTEEVENTS\",\n",
    "    split=\"train\"\n",
    ")\n",
    "\n",
    "# Inspect columns\n",
    "print(\"Columns:\", notes.column_names)\n",
    "\n",
    "# Filter discharge summaries\n",
    "discharge_notes = notes.filter(\n",
    "    lambda x: x[\"CATEGORY\"] == \"Discharge summary\"\n",
    ")\n",
    "\n",
    "print(\"Total discharge summaries:\", len(discharge_notes))\n",
    "\n",
    "# Convert small sample to pandas for inspection\n",
    "df = discharge_notes.select(range(5)).to_pandas()\n",
    "\n",
    "# Print sample text\n",
    "for i, text in enumerate(df[\"TEXT\"]):\n",
    "    print(f\"\\n--- Discharge Summary {i+1} ---\\n\")\n",
    "    print(text[:2000])\n"
   ]
  }
 ],
 "metadata": {
  "language_info": {
   "name": "python"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}