Learning Diffusion Policies from Demonstrations For Compliant Contact-rich Manipulation