DiffusionAct: Controllable Diffusion Autoencoder for One-shot Face Reenactment