SFDiff: Diffusion Model with Sufficient Spatial-Fourier Frequency Information Interaction for Low-Light Image Enhancement