Recursive Graph Pattern Matching With Magic Sets and Global Search Plans.pdfVIP

Recursive Graph Pattern Matching With Magic Sets and Global Search Plans ´1 ´ ´ 2 ´ ´2 Gergely Varro , Akos Horvath , and Daniel Varro 1 Department of Computer Science and Information Theory Budapest University of Technology and Economics gervarro@cs.bme.hu 2 Department of Measurement and Information Systems Budapest University of Technology and Economics [ahorvath,varro]@mit.bme.hu Abstract. We present core data structures and algorithms for matching graph patterns with general recursion. Our approach uses magic sets, a well-known technique from deductive databases, which combines fixpoint-based bottom-up query evaluation with top-down handling of input parameters. Furthermore, this technique is enhanced with the global search plans, thus non-recursive calls are always flattened before elementary pattern matching operations are initiated in order to improve performance. Our approach is exemplified using VIATRA2. 1 Introduction Graph transformation (GT) [1] is a frequently used means to capture model transfor- mations in the context of model-driven software development. Graph transformation rules provide a declarative, rule and pattern-based language for specifying both inter- language and intra-language model manipulations for model analysis, refactoring or simulation. GT rules consist of a left-hand side (LHS) and a right-hand side (RHS) graph pat- tern. The LHS specifies contextual conditions which should hold as a precondition for applying the rule, which is checked by graph pattern matching. Then the model is ma- nipulated by calculating the difference of the RHS and the LHS in the