SPARQL ̶ RW : Transparent Query Access over Mapped RDF Data Sources Kostantinos Makris 1 Nikos Bikakis 2,3 Nektarios Gioldasis 1 Stavros Christodoulakis 1

EDBT 2012 | 15th International Conference on Extending Database Technology | Berlin | Germany

Sparql

ReWritingSRW

The SPARQL2XQuery Framework

Bridging the Gab between the XML and the Semantic Web Worlds.

· XML Schema to OWL Transformation (XS2OWL plug-in)

· Mapping Specification & Generation between OWL - RDF/S and XML Schemas

· SPARQL to XQuery Query Translation

· XML ⎯ RDF Data Transformations

Family Framework

X

2

Sparql

XQuery

XS

Source Target

Motivating Example

1 Technical University of Crete, Greece 2

National Technical University of Athens, Greece

3 "Athena" Research Center, Greece

Query ReWriting

Intro

· Describes mappings between OWL/RDF-S ontology schemas· Defines and supports all the possible mapping types which can be exploited by the SPARQL query rewriting process.· The supported mappings are highly dependent to the SPARQL expressiveness.

Grammar

· Four basic notions: ‒ Class Expression ‒ Object Property Expression ‒ Datatype Property Expression ‒ Instance Expression· Expressions are defined using several operators: union (⨆), intersection (⨅), composition (○), inverse (—), domain/range restrictions, etc.· Supports N:M cardinality mappings using equivalence or subsumption relationships.

SemanticsFormal semantics based on Description Logics.

Mapping Model

The SPARQL‒RW Framework

The SPARQL-RW (SPARQL ReWriting) Framework provides a generic method for

SPARQL query rewriting, with respect to a set of predefined mappings between

ontology schemas

Key Features

· Mapping Model ‒ Specifies mapping types which can be

exploited by the SPARQL query rewriting process

‒ Formally described using Description Logics ‒ Supports Flexible & Rich Mappings

· Query Rewriting Method ‒ Based on a set of predefined mappings

between ontologies ‒ Semantics Preserving

· Java System Prototype

Architecture

Problem Definition

Let a source ontology OS, a target ontology OT

and a set of mappings M between OS and OT. Our framework takes as input a SPARQL query QS expressed over OS, and rewrites it to a semantically correspondent SPARQL query QT (expressed over OT) w.r.t. M.

Intro

Overview

· The SPARQL query rewriting process lies in the query’s graph pattern rewriting. · For each Basic Graph Pattern, rewrite each triple pattern a) by Predicate b) by Object c) by Subject.· The rewriting is independent of the query type (i.e., Select, Ask), the SPARQL solution sequence modifiers (i.e., Order By, Distinct) and the SPARQL algebra operators (i.e., Union, Optional).· Triple Patterns Types ‒ Data Triple Patterns: Deal with data and not schema info. ‒ Schema Triple Patterns: Deal with schema info. Contain RDF/RDFS/OWL terms.

Semantics Preserving Rewriting

Let DS and DT be the RDF datasets of a source and a target ontology, respectively. Similarly, let DU be the RDF dataset which is produced by merging the DS and DT datasets using a set of mappings M.

Definition (Semantics Preserving Rewriting). Let tp be a triple pattern and rp the graph pattern resulted from one step rewriting of tp w.r.t. a mapping μ M. The rewriting step performed for tp, w.r.t. the mapping μ, is semantics preserving iff the evaluation of tp and the evaluation of rp over DU, preserve the semantics of mapping μ.

In other words, let V be the common variable set between tp and rp. The relationship R (i.e., ≡, ⊑, ⊒) that holds for the mapping used in the rewriting step, should also hold between [[tp]]DU and [[rp]]DU

projected on V.

πV ([[tp]]DU) R πV ([[rp]]DU), where R { ≡, ⊑, ⊒ }

SPARQL QueriesUsing DBpedia 3.2 Vocabulary

Select ?x { ?x dbpprop-el:όνομα "Christos Papadimitriou" }

DBpedia

v. 3.7

DBLP

ACM

YAGO

IEEE

Freebase

Greek

DBpedia

DBpedia

v. 3.2

Initial TriplesRewritten Triples by

Predicate PartRewritten Triples by

Object Part

Predicate Mappings

Object Mappings

Subject Mappings Rewritten Triples by

Subject Part (Final Triples)

Mappings(RDF)

SPARQL

Qt

Qs

RDF Data

Target Ontology

SPARQL‒RW

RDF

RDF Data

Source Ontology

Mapping Type Determinator

Mapping Parser

Query Analyzer & Composer

Results VisualizerGraph Pattern Rewriter

Rewriting Rules& Axioms

FILTER Expr. Rewriter

Triple Pattern Rewriter

Triple Pattern Type Determinator

Subject Rewriter

Object Rewriter

Predicate Rewriter