Add Spanish notebooks (chat_vision, chat_pdf_images)

notebooks/Spanish/chat_pdf_images.ipynb

@@ -0,0 +1,182 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "id": "fd77e3b8",
+   "metadata": {},
+   "source": [
+    "# Chatea con imágenes de páginas PDF\n",
+    "\n",
+    "**Si estás buscando la aplicación web, revisa la carpeta src/.** \n",
+    "\n",
+    "Este notebook demuestra cómo convertir páginas de PDF a imágenes y enviarlas a un modelo de visión para inferencia"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "e5eb545b",
+   "metadata": {},
+   "source": [
+    "## Autentícate en OpenAI\n",
+    "\n",
+    "El siguiente código se conecta a OpenAI, ya sea usando una cuenta de Azure OpenAI, modelos de GitHub, o modelo local de Ollama. Mira el README para instrucciones sobre cómo configurar el archivo `.env`."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "6ae3a4d3",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import os\n",
+    "\n",
+    "import azure.identity\n",
+    "import openai\n",
+    "from dotenv import load_dotenv\n",
+    "\n",
+    "load_dotenv(\".env\", override=True)\n",
+    "\n",
+    "openai_host = os.getenv(\"OPENAI_HOST\", \"github\")\n",
+    "\n",
+    "if openai_host == \"github\":\n",
+    "    print(\"Usando GitHub Models con GITHUB_TOKEN como clave\")\n",
+    "    openai_client = openai.OpenAI(\n",
+    "        api_key=os.environ[\"GITHUB_TOKEN\"],\n",
+    "        base_url=\"https://models.github.ai/inference\",\n",
+    "    )\n",
+    "    model_name = os.getenv(\"OPENAI_MODEL\", \"openai/gpt-4o\")\n",
+    "elif openai_host == \"local\":\n",
+    "    print(\"Usando API local compatible con OpenAI sin clave\")\n",
+    "    openai_client = openai.OpenAI(api_key=\"no-key-required\", base_url=os.environ[\"LOCAL_OPENAI_ENDPOINT\"])\n",
+    "    model_name = os.getenv(\"OPENAI_MODEL\", \"gpt-4o\")\n",
+    "elif openai_host == \"azure\" and os.getenv(\"AZURE_OPENAI_KEY_FOR_CHATVISION\"):\n",
+    "    # Autenticación usando una clave de API de Azure OpenAI\n",
+    "    # Esto generalmente no se recomienda, pero se proporciona por conveniencia\n",
+    "    print(\"Usando Azure OpenAI con clave\")\n",
+    "    openai_client = openai.OpenAI(\n",
+    "        base_url=os.environ[\"AZURE_OPENAI_ENDPOINT\"] + \"/openai/v1/\",\n",
+    "        api_key=os.environ[\"AZURE_OPENAI_KEY_FOR_CHATVISION\"],\n",
+    "    )\n",
+    "    # Esto es en realidad el nombre del deployment, no el nombre del modelo\n",
+    "    model_name = os.getenv(\"OPENAI_MODEL\", \"gpt-4o\")\n",
+    "elif openai_host == \"azure\" and os.getenv(\"AZURE_OPENAI_ENDPOINT\"):\n",
+    "    tenant_id = os.environ[\"AZURE_TENANT_ID\"]\n",
+    "    print(\"Usando Azure OpenAI con credencial de Azure Developer CLI para tenant id\", tenant_id)\n",
+    "    default_credential = azure.identity.AzureDeveloperCliCredential(tenant_id=tenant_id)\n",
+    "    token_provider = azure.identity.get_bearer_token_provider(\n",
+    "        default_credential, \"https://cognitiveservices.azure.com/.default\"\n",
+    "    )\n",
+    "    openai_client = openai.OpenAI(\n",
+    "        base_url=os.environ[\"AZURE_OPENAI_ENDPOINT\"] + \"/openai/v1/\",\n",
+    "        api_key=token_provider,\n",
+    "    )\n",
+    "    # Esto es en realidad el nombre del deployment, no el nombre del modelo\n",
+    "    model_name = os.getenv(\"OPENAI_MODEL\", \"gpt-4o\")\n",
+    "\n",
+    "print(f\"Usando modelo {model_name}\")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "74df1ca5",
+   "metadata": {},
+   "source": [
+    "## Convierte PDFs a imágenes"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "e3664e1d",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "%pip install Pillow PyMuPDF"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "c8c56b19",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import pymupdf\n",
+    "from PIL import Image\n",
+    "\n",
+    "filename = \"../plants.pdf\"\n",
+    "doc = pymupdf.open(filename)\n",
+    "for i in range(doc.page_count):\n",
+    "    doc = pymupdf.open(filename)\n",
+    "    page = doc.load_page(i)\n",
+    "    pix = page.get_pixmap()\n",
+    "    original_img = Image.frombytes(\"RGB\", [pix.width, pix.height], pix.samples)\n",
+    "    original_img.save(f\"page_{i}.png\")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "f822fb8f",
+   "metadata": {},
+   "source": [
+    "## Envía imágenes al modelo de visión"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "1bdaa995",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import base64\n",
+    "\n",
+    "\n",
+    "def open_image_as_base64(filename):\n",
+    "    with open(filename, \"rb\") as image_file:\n",
+    "        image_data = image_file.read()\n",
+    "    image_base64 = base64.b64encode(image_data).decode(\"utf-8\")\n",
+    "    return f\"data:image/png;base64,{image_base64}\""
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "4de5dca7",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "user_content = [{\"text\": \"¿Qué plantas están listadas en estas páginas?\", \"type\": \"text\"}]\n",
+    "# Procesa solo las primeras páginas, ya que procesar todas las páginas (doc.page_count) es lento\n",
+    "for i in range(3):\n",
+    "    user_content.append({\"image_url\": {\"url\": open_image_as_base64(f\"../page_{i}.png\")}, \"type\": \"image_url\"})\n",
+    "\n",
+    "response = openai_client.chat.completions.create(model=model_name, messages=[{\"role\": \"user\", \"content\": user_content}])\n",
+    "\n",
+    "print(response.choices[0].message.content)"
+   ]
+  }
+ ],
+ "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": "3.12.11"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}