{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "36832aa6",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1.2.2\n"
     ]
    }
   ],
   "source": [
    "import duckdb\n",
    "print(duckdb.__version__)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "c357d895",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<duckdb.duckdb.DuckDBPyConnection at 0x1cda135ac30>"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "conn.execute(\"INSTALL vss; LOAD vss;\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "4a7a3ffb",
   "metadata": {},
   "outputs": [
    {
     "ename": "BinderException",
     "evalue": "Binder Error: HNSW indexes can only be created in in-memory databases, or when the configuration option 'hnsw_enable_experimental_persistence' is set to true.",
     "output_type": "error",
     "traceback": [
      "\u001b[31m---------------------------------------------------------------------------\u001b[39m",
      "\u001b[31mBinderException\u001b[39m                           Traceback (most recent call last)",
      "\u001b[36mCell\u001b[39m\u001b[36m \u001b[39m\u001b[32mIn[13]\u001b[39m\u001b[32m, line 38\u001b[39m\n\u001b[32m     35\u001b[39m     conn.execute(\u001b[33m\"\u001b[39m\u001b[33mINSERT INTO textos VALUES (?, ?, ?)\u001b[39m\u001b[33m\"\u001b[39m, (i, texto, vector))\n\u001b[32m     37\u001b[39m \u001b[38;5;66;03m# Crear el índice HNSW en la columna embedding\u001b[39;00m\n\u001b[32m---> \u001b[39m\u001b[32m38\u001b[39m \u001b[43mconn\u001b[49m\u001b[43m.\u001b[49m\u001b[43mexecute\u001b[49m\u001b[43m(\u001b[49m\u001b[33;43m\"\u001b[39;49m\u001b[33;43mCREATE INDEX vss_idx ON textos USING HNSW (embedding);\u001b[39;49m\u001b[33;43m\"\u001b[39;49m\u001b[43m)\u001b[49m\n\u001b[32m     40\u001b[39m \u001b[38;5;66;03m# Realizar una búsqueda de los 3 textos más similares a una consulta\u001b[39;00m\n\u001b[32m     41\u001b[39m consulta = \u001b[33m\"\u001b[39m\u001b[33minteligencia artificial en Python\u001b[39m\u001b[33m\"\u001b[39m\n",
      "\u001b[31mBinderException\u001b[39m: Binder Error: HNSW indexes can only be created in in-memory databases, or when the configuration option 'hnsw_enable_experimental_persistence' is set to true."
     ]
    }
   ],
   "source": [
    "import duckdb\n",
    "import numpy as np\n",
    "from sentence_transformers import SentenceTransformer\n",
    "\n",
    "# Conectar a la base de datos\n",
    "conn = duckdb.connect(\"textos_vss.duckdb\")\n",
    "\n",
    "# Instalar y cargar la extensión VSS\n",
    "conn.execute(\"INSTALL vss;\")\n",
    "conn.execute(\"LOAD vss;\")\n",
    "\n",
    "# Modelo de embeddings\n",
    "model = SentenceTransformer(\"all-MiniLM-L6-v2\")\n",
    "\n",
    "# Crear la tabla con una columna de embeddings de tipo FLOAT[]\n",
    "conn.execute(\"\"\"\n",
    "CREATE TABLE IF NOT EXISTS textos (\n",
    "    id INTEGER PRIMARY KEY,\n",
    "    texto TEXT,\n",
    "    embedding FLOAT[]\n",
    ");\n",
    "\"\"\")\n",
    "\n",
    "# Insertar textos y sus embeddings\n",
    "extractos = [\n",
    "    \"El rápido zorro marrón salta sobre el perro perezoso.\",\n",
    "    \"La inteligencia artificial está revolucionando el mundo.\",\n",
    "    \"DuckDB permite análisis embebidos rápidos en Python.\",\n",
    "    \"Los modelos de lenguaje pueden entender consultas humanas.\",\n",
    "    \"Python es un lenguaje poderoso para ciencia de datos.\"\n",
    "]\n",
    "\n",
    "for i, texto in enumerate(extractos):\n",
    "    vector = model.encode(texto).astype(np.float32).tolist()\n",
    "    conn.execute(\"INSERT INTO textos VALUES (?, ?, ?)\", (i, texto, vector))\n",
    "\n",
    "# Crear el índice HNSW en la columna embedding\n",
    "conn.execute(\"CREATE INDEX vss_idx ON textos USING HNSW (embedding);\")\n",
    "\n",
    "# Realizar una búsqueda de los 3 textos más similares a una consulta\n",
    "consulta = \"inteligencia artificial en Python\"\n",
    "vector_consulta = model.encode(consulta).astype(np.float32).tolist()\n",
    "resultados = conn.execute(\"\"\"\n",
    "    SELECT id, texto, array_distance(embedding, ?) AS distancia\n",
    "    FROM textos\n",
    "    ORDER BY distancia ASC\n",
    "    LIMIT 3;\n",
    "\"\"\", (vector_consulta,)).fetchall()\n",
    "\n",
    "# Mostrar los resultados\n",
    "for id_, texto, distancia in resultados:\n",
    "    print(f\"[ID: {id_}] Distancia: {distancia:.4f} → {texto}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "2f6059bd",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<duckdb.duckdb.DuckDBPyConnection at 0x1cda32d9730>"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "conn = duckdb.connect(\"textos_vss.duckdb\")\n",
    "conn.execute(\"SET hnsw_enable_experimental_persistence=true;\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "449bebc1",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<duckdb.duckdb.DuckDBPyConnection at 0x1cda322a5b0>"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "conn.execute(\"DELETE FROM textos;\")\n"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": ".venv",
   "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": "3.11.11"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}