app.py

import os
import uuid
from typing import List, Tuple, Optional, Union
from PIL import Image
import google.generativeai as genai
import gradio as gr
from dotenv import load_dotenv

# Cargar las variables de entorno desde el archivo .env
load_dotenv()

# Configuración de claves
GOOGLE_API_KEY = os.getenv("GOOGLE_API_KEY")
if not GOOGLE_API_KEY:
    raise ValueError("GOOGLE_API_KEY is not set in environment variables.")

genai.configure(api_key=GOOGLE_API_KEY)

# Configuración general
TITLE = """<h1 align="center">Gemini Playground ✨</h1>"""
SUBTITLE = """<h2 align="center">Play with Gemini Pro and Gemini Pro Vision</h2>"""

IMAGE_CACHE_DIRECTORY = "/tmp"
IMAGE_WIDTH = 512

# Función para preprocesar imágenes
def preprocess_image(image: Image.Image) -> Optional[Image.Image]:
    if image:
        image_height = int(image.height * IMAGE_WIDTH / image.width)
        return image.resize((IMAGE_WIDTH, image_height))

# Función para guardar imágenes en caché
def cache_pil_image(image: Image.Image) -> str:
    image_filename = f"{uuid.uuid4()}.jpeg"
    os.makedirs(IMAGE_CACHE_DIRECTORY, exist_ok=True)
    image_path = os.path.join(IMAGE_CACHE_DIRECTORY, image_filename)
    image.save(image_path, "JPEG")
    return image_path

# Pestaña 1: Chatbot de solo texto con historial
def bot_response(
    model_choice: str,
    system_instruction: str,
    text_prompt: str,
    chatbot: List[Tuple[str, str]],
) -> Tuple[List[Tuple[str, str]], str]:
    if not text_prompt.strip():
        return chatbot, "Por favor, escribe un mensaje válido."

    model = genai.GenerativeModel(
        model_name=model_choice,
        generation_config={
            "temperature": 1,
            "top_p": 0.95,
            "top_k": 40,
            "max_output_tokens": 8192,
        },
    )
    chat = model.start_chat(history=chatbot)
    chat.system_instruction = system_instruction

    response = chat.send_message(text_prompt)
    response.resolve()

    chatbot.append((text_prompt, response.text))
    return chatbot, ""

# Pestaña 2: Chatbot avanzado con imágenes
def upload(files: Optional[List[gr.File]], chatbot: List[Tuple[str, str]]) -> List[Tuple[str, str]]:
    if files:
        for file in files:
            image = Image.open(file.name).convert("RGB")
            image_preview = preprocess_image(image)
            if image_preview:
                image_path = cache_pil_image(image_preview)
                chatbot.append((f"Uploaded image: {image_path}", "Imagen cargada correctamente."))
    return chatbot

def advanced_response(
    text_prompt: str,
    files: Optional[List[gr.File]],
    model_choice: str,
    system_instruction: str,
    chatbot: List[Tuple[str, str]],
):
    if not text_prompt.strip() and not files:
        chatbot.append(("", "Por favor, proporciona un mensaje o sube una imagen."))
        yield chatbot
        return

    model = genai.GenerativeModel(
        model_name=model_choice,
        generation_config={
            "temperature": 0.7,
            "max_output_tokens": 8192,
            "top_k": 10,
            "top_p": 0.9,
        },
    )
    chat = model.start_chat(history=chatbot)
    chat.system_instruction = system_instruction

    if text_prompt:
        chatbot.append((text_prompt, ""))
    if files:
        images = [cache_pil_image(preprocess_image(Image.open(file.name))) for file in files]
        chatbot.append((f"Uploaded images: {', '.join(images)}", ""))

    response = chat.send_message(text_prompt)
    response.resolve()

    chatbot[-1] = (chatbot[-1][0], response.text)
    yield chatbot

# Construcción de la interfaz
def build_interface():
    with gr.Blocks() as demo:
        gr.HTML(TITLE)
        gr.HTML(SUBTITLE)

        with gr.Tab("Pestaña 1: Chatbot Texto"):
            model_dropdown_1 = gr.Dropdown(
                choices=["gemini-1.5-flash", "gemini-2.0-flash-exp", "gemini-1.5-pro"],
                value="gemini-1.5-flash",
                label="Selecciona el modelo",
            )
            chatbot_1 = gr.Chatbot(label="Gemini", height=300)
            system_instruction_1 = gr.Textbox(
                placeholder="Escribe una instrucción para el sistema...",
                label="Instrucción del sistema",
                value="You are an assistant.",
            )

            with gr.Row():
                text_input_1 = gr.Textbox(placeholder="Escribe un mensaje...", show_label=False)
                run_button_1 = gr.Button(value="Enviar", variant="primary")

            run_button_1.click(
                fn=bot_response,
                inputs=[model_dropdown_1, system_instruction_1, text_input_1, chatbot_1],
                outputs=[chatbot_1, text_input_1],
            )

        with gr.Tab("Pestaña 2: Chatbot con Imágenes"):
            model_dropdown_2 = gr.Dropdown(
                choices=["gemini-1.5-flash", "gemini-2.0-flash-exp", "gemini-1.5-pro"],
                value="gemini-1.5-flash",
                label="Selecciona el modelo",
            )
            chatbot_2 = gr.Chatbot(label="Gemini", height=300)
            system_instruction_2 = gr.Textbox(
                placeholder="Escribe una instrucción para el sistema...",
                label="Instrucción del sistema",
            )

            with gr.Row():
                text_input_2 = gr.Textbox(placeholder="Mensaje o descripción...", show_label=False)
                upload_button = gr.UploadButton(label="Subir Imágenes", file_count="multiple", file_types=["image"])
                run_button_2 = gr.Button(value="Ejecutar", variant="primary")

            run_button_2.click(
                fn=advanced_response,
                inputs=[text_input_2, upload_button, model_dropdown_2, system_instruction_2, chatbot_2],
                outputs=[chatbot_2],
            )
            upload_button.upload(
                fn=upload,
                inputs=[upload_button, chatbot_2],
                outputs=[chatbot_2],
            )

    return demo

if __name__ == "__main__":
    demo = build_interface()
    demo.launch(debug=True)