Magic shapes for SHACL validation

A key prerequisite for the successful adoption of Shapes Constraint Language (SHACL)---the W3C standardized constraint language RDF graphs---is availability automated tools that efficiently validate targeted constraints (known as shapes graphs ) over possibly very large graphs. There are already significant efforts to produce optimized engines SHACL validation, but they focus on restricted fragments SHACL. For unrestricted SHACL, is with recursion and negation, there no validator beyond a proof-of-concept prototype, existing techniques inherently incompatible goal-driven approaches being pursued by validators. Instead require global computation entire data graph not only computationally costly, also brittle, can easily result in validation failures due conflicts irrelevant targets. To address these challenges, we present 'magic' transformation---based Magic Sets known from Logic Programming---that transforms S into new S' whose considers relevant neighbourhood nodes. The equivalent whenever between data, case fails target, may still admit lazy, target-oriented validation. We implement algorithm run preliminary experiments, showing our approach be stepping stone towards validators full it significantly improve performance prototype currently supports negation.