app.py

# app.py
import gradio as gr
import torch
import jax
import jax.numpy as jnp
import numpy as np
from PIL import Image
import pickle
import logging
from huggingface_hub import hf_hub_download
from diffusers import StableDiffusionXLImg2ImgPipeline
from transformers import DPTImageProcessor, DPTForDepthEstimation
from model import build_thera
from utils import make_grid, interpolate_grid

# Configuração de logging
logging.basicConfig(
    level=logging.INFO,
    format='%(asctime)s - %(levelname)s - %(message)s',
    handlers=[logging.FileHandler("processing.log"), logging.StreamHandler()]
)
logger = logging.getLogger(__name__)

# Configurações
JAX_DEVICE = jax.devices("cpu")[0]
TORCH_DEVICE = "cpu"


def load_thera_model(repo_id: str, filename: str):
    """Carrega modelo com verificação de segurança"""
    try:
        model_path = hf_hub_download(repo_id=repo_id, filename=filename)
        with open(model_path, 'rb') as fh:
            checkpoint = pickle.load(fh)
        return build_thera(3, checkpoint['backbone'], checkpoint['size']), checkpoint['model']
    except Exception as e:
        logger.error(f"Erro ao carregar modelo: {str(e)}")
        raise


# Inicialização dos modelos
try:
    logger.info("Carregando modelos...")
    model_edsr, params_edsr = load_thera_model("prs-eth/thera-edsr-pro", "model.pkl")
    pipe = StableDiffusionXLImg2ImgPipeline.from_pretrained(
        "stabilityai/stable-diffusion-xl-base-1.0",
        torch_dtype=torch.float32
    ).to(TORCH_DEVICE)
    pipe.load_lora_weights("KappaNeuro/bas-relief", weight_name="BAS-RELIEF.safetensors")
    feature_extractor = DPTImageProcessor.from_pretrained("Intel/dpt-large")
    depth_model = DPTForDepthEstimation.from_pretrained("Intel/dpt-large").to(TORCH_DEVICE)
except Exception as e:
    logger.error(f"Falha na inicialização: {str(e)}")
    raise


def adjust_size(original: int, scale: float) -> int:
    """Ajuste de tamanho com limites seguros"""
    scaled = int(original * scale)
    adjusted = (scaled // 8) * 8  # Divisível por 8
    return max(32, adjusted)  # Mínimo absoluto


def full_pipeline(image: Image.Image, prompt: str, scale_factor: float = 2.0):
    """Pipeline completo com tratamento robusto"""
    try:
        # Pré-processamento
        image = image.convert("RGB")
        orig_w, orig_h = image.size

        # Cálculo do tamanho alvo
        new_h = adjust_size(orig_h, scale_factor)
        new_w = adjust_size(orig_w, scale_factor)
        logger.info(f"Redimensionando: {orig_h}x{orig_w} → {new_h}x{new_w}")

        # Gerar grid de coordenadas
        coords = make_grid((new_h, new_w))
        logger.debug(f"Dimensões do grid: {coords.shape}")

        # Verificação crítica
        if coords.shape[1:3] != (new_h, new_w):
            raise ValueError(f"Grid incorreto: {coords.shape[1:3]} vs ({new_h}, {new_w})")

        # Super-resolução
        source = jnp.array(image).astype(jnp.float32) / 255.0
        source = source[jnp.newaxis, ...]  # Adicionar batch

        t = jnp.array([1.0 / (scale_factor ** 2)], dtype=jnp.float32)
        upscaled = model_edsr.apply(params_edsr, source, t, (new_h, new_w))

        # Pós-processamento
        upscaled_img = Image.fromarray((np.array(upscaled[0]) * 255).astype(np.uint8))

        # Bas-Relief
        result = pipe(
            prompt=f"BAS-RELIEF {prompt}, ultra detailed, 8K resolution",
            image=upscaled_img,
            strength=0.7,
            num_inference_steps=30
        )
        bas_relief = result.images[0]

        # Mapa de profundidade
        inputs = feature_extractor(bas_relief, return_tensors="pt").to(TORCH_DEVICE)
        with torch.no_grad():
            depth = depth_model(**inputs).predicted_depth

        depth_map = torch.nn.functional.interpolate(
            depth.unsqueeze(1),
            size=bas_relief.size[::-1],
            mode="bicubic"
        ).squeeze().cpu().numpy()

        # Normalização
        depth_min = depth_map.min()
        depth_max = depth_map.max()
        depth_normalized = (depth_map - depth_min) / (depth_max - depth_min + 1e-8)
        depth_img = Image.fromarray((depth_normalized * 255).astype(np.uint8))

        return upscaled_img, bas_relief, depth_img

    except Exception as e:
        logger.error(f"ERRO NO PIPELINE: {str(e)}", exc_info=True)
        raise gr.Error(f"Processamento falhou: {str(e)}")


# Interface
with gr.Blocks(title="SuperRes+BasRelief", theme=gr.themes.Default()) as app:
    gr.Markdown("# 🖼️ Super Resolução + 🗿 Bas-Relief + 🗺️ Mapa de Profundidade")

    with gr.Row():
        with gr.Column():
            img_input = gr.Image(label="Imagem de Entrada", type="pil")
            prompt = gr.Textbox(
                label="Descrição do Relevo",
                value="Ainsanely detailed and complex engraving relief, ultra-high definition",
                placeholder="Descreva o estilo desejado..."
            )
            scale = gr.Slider(1.0, 4.0, value=2.0, label="Fator de Escala")
            btn = gr.Button("Processar Imagem", variant="primary")

        with gr.Column():
            gr.Markdown("## Resultados")
            with gr.Tabs():
                with gr.TabItem("Super Resolução"):
                    upscaled_output = gr.Image(label="Resultado Super Resolução")
                with gr.TabItem("Bas-Relief"):
                    basrelief_output = gr.Image(label="Relevo Gerado")
                with gr.TabItem("Profundidade"):
                    depth_output = gr.Image(label="Mapa de Profundidade")

    btn.click(
        full_pipeline,
        inputs=[img_input, prompt, scale],
        outputs=[upscaled_output, basrelief_output, depth_output]
    )

if __name__ == "__main__":
    app.launch(server_name="0.0.0.0", server_port=7860)