A Framework for Goal-Directed Bottom-Up Evaluation of Functional Logic Programs