Given the facilities we have here, I’d go with short alt text and a longer description in the body. That way, screen reader users know what’s in the image, but everybody gets the explanation and context.