Ontop: Answering SPARQL Queries over Relational Databases

1. Ontop: Answering SPARQL Queries over Relational Databases Guohui Xiao Faculty of Computer Science, Free University of Bozen-Bolzano, Italy Free University of Bozen-Bolzano February 12, 2016 Stanford University, CA, USA

2. Introduction Overview of Ontop SPARQL Query Answering in Ontop Use Cases Recent Progresses and Future About Me • Guohui Xiao, PhD • Assistant Professor at KRDB Research Centre for Knowledge and Data, Free University of Bozen-Bolzano, Italy • Educations PhD in Computer Science, Vienna University of Technology, Austria MSc and BSc in Mathematics, Peking University, China • Research interests: Artiﬁcial intelligence, Knowledge representation Description logics, Ontology, Semantic Web Ontology-based Data Access Implementation and Optimization of reasoning systems • Ontop team leader • Current project: Optique (Scalable End-user Access to Big Data), EU FP7 G. Xiao (FUB) Ontop: Answering SPARQL Queries over Relational Databases 1/56

3. Introduction Overview of Ontop SPARQL Query Answering in Ontop Use Cases Recent Progresses and Future Outline 1 Introduction 2 Overview of Ontop 3 SPARQL Query Answering in Ontop 4 Use Cases 5 Recent Progresses and Future G. Xiao (FUB) Ontop: Answering SPARQL Queries over Relational Databases 2/56

5. Introduction Overview of Ontop SPARQL Query Answering in Ontop Use Cases Recent Progresses and Future We are Living in the Era of Big Data Data Never Sleeps 2.0 G. Xiao (FUB) Ontop: Answering SPARQL Queries over Relational Databases 4/56

6. Introduction Overview of Ontop SPARQL Query Answering in Ontop Use Cases Recent Progresses and Future The Problem: information access G. Xiao (FUB) Ontop: Answering SPARQL Queries over Relational Databases 5/56

7. Introduction Overview of Ontop SPARQL Query Answering in Ontop Use Cases Recent Progresses and Future The Problem: information access How to formulate the right question to obtain the right answer in the ocean of Big Data. G. Xiao (FUB) Ontop: Answering SPARQL Queries over Relational Databases 5/56

8. Introduction Overview of Ontop SPARQL Query Answering in Ontop Use Cases Recent Progresses and Future How much time is spent searching for data? Engineers in industry spend a signiﬁcant amount of their time searching for data that they require for their core tasks. For example, in the oil&gas industry, 30–70% of engineers’ time is spent looking for data and assessing its quality (Crompton, 2008). G. Xiao (FUB) Ontop: Answering SPARQL Queries over Relational Databases 6/56

9. Introduction Overview of Ontop SPARQL Query Answering in Ontop Use Cases Recent Progresses and Future Example: Statoil Exploration Experts in geology and geophysics develop stratigraphic models of unexplored areas on the basis of data acquired from previous operations at nearby locations. G. Xiao (FUB) Ontop: Answering SPARQL Queries over Relational Databases 7/56

10. Introduction Overview of Ontop SPARQL Query Answering in Ontop Use Cases Recent Progresses and Future Example: Statoil Exploration Experts in geology and geophysics develop stratigraphic models of unexplored areas on the basis of data acquired from previous operations at nearby locations. Facts: • 1,000 TB of relational data • using diverse schemata • spread over 2,000 tables, over multiple individual data bases Data Access for Exploration: • 900 experts in Statoil Exploration. G. Xiao (FUB) Ontop: Answering SPARQL Queries over Relational Databases 7/56

11. Introduction Overview of Ontop SPARQL Query Answering in Ontop Use Cases Recent Progresses and Future How much time/money is spent searching for data? A user query at Statoil Show all norwegian wellbores with some additional attributes (wellbore id, .....................). Limit to all wellbores with ... and show attributes like ............................................... Limit to all wellbores with ... in ................. and show key attributes in a table. After connecting to ... we could for instance limit further to cores in ... with ...... and where it is larger than a given value, for instance ..... We could also ﬁnd out whether there are cores in ..... which are not stored in .... (based on .....) and where there could be .......... value. Some of the missing data we possibly own, other not. G. Xiao (FUB) Ontop: Answering SPARQL Queries over Relational Databases 8/56

12. Introduction Overview of Ontop SPARQL Query Answering in Ontop Use Cases Recent Progresses and Future How much time/money is spent searching for data? A user query at Statoil Show all norwegian wellbores with some additional attributes (wellbore id, .....................). Limit to all wellbores with ... and show attributes like ............................................... Limit to all wellbores with ... in ................. and show key attributes in a table. After connecting to ... we could for instance limit further to cores in ... with ...... and where it is larger than a given value, for instance ..... We could also ﬁnd out whether there are cores in ..... which are not stored in .... (based on .....) and where there could be .......... value. Some of the missing data we possibly own, other not. SELECT [...] FROM db_name.table1 table1, db_name.table2 table2a, db_name.table2 table2b, db_name.table3 table3a, db_name.table3 table3b, db_name.table3 table3c, db_name.table3 table3d, db_name.table4 table4a, db_name.table4 table4b, db_name.table4 table4c, db_name.table4 table4d, db_name.table4 table4e, db_name.table4 table4f, db_name.table5 table5a, db_name.table5 table5b, db_name.table6 table6a, db_name.table6 table6b, db_name.table7 table7a, db_name.table7 table7b, db_name.table8 table8, db_name.table9 table9, db_name.table10 table10a, db_name.table10 table10b, db_name.table10 table10c, db_name.table11 table11, db_name.table12 table12, db_name.table13 table13, db_name.table14 table14, db_name.table15 table15, db_name.table16 table16 WHERE [...] table2a.attr1=‘keyword’ AND table3a.attr2=table10c.attr1 AND table3a.attr6=table6a.attr3 AND table3a.attr9=‘keyword’ AND table4a.attr10 IN (‘keyword’) AND table4a.attr1 IN (‘keyword’) AND table5a.kinds=table4a.attr13 AND table5b.kinds=table4c.attr74 AND table5b.name=‘keyword’ AND (table6a.attr19=table10c.attr17 OR (table6a.attr2 IS NULL AND table10c.attr4 IS NULL)) AND table6a.attr14=table5b.attr14 AND table6a.attr2=‘keyword’ AND (table6b.attr14=table10c.attr8 OR (table6b.attr4 IS NULL AND table10c.attr7 IS NULL)) AND table6b.attr19=table5a.attr55 AND table6b.attr2=‘keyword’ AND table7a.attr19=table2b.attr19 AND table7a.attr17=table15.attr19 AND table4b.attr11=‘keyword’ AND table8.attr19=table7a.attr80 AND table8.attr19=table13.attr20 AND table8.attr4=‘keyword’ AND table9.attr10=table16.attr11 AND table3b.attr19=table10c.attr18 AND table3b.attr22=table12.attr63 AND table3b.attr66=‘keyword’ AND table10a.attr54=table7a.attr8 AND table10a.attr70=table10c.attr10 AND table10a.attr16=table4d.attr11 AND table4c.attr99=‘keyword’ AND table4c.attr1=‘keyword’ AND table11.attr10=table5a.attr10 AND table11.attr40=‘keyword’ AND table11.attr50=‘keyword’ AND table2b.attr1=table1.attr8 AND table2b.attr9 IN (‘keyword’) AND table2b.attr2 LIKE ‘keyword’% AND table12.attr9 IN (‘keyword’) AND table7b.attr1=table2a.attr10 AND table3c.attr13=table10c.attr1 AND table3c.attr10=table6b.attr20 AND table3c.attr13=‘keyword’ AND table10b.attr16=table10a.attr7 AND table10b.attr11=table7b.attr8 AND table10b.attr13=table4b.attr89 AND table13.attr1=table2b.attr10 AND table13.attr20=’‘keyword’’ AND table13.attr15=‘keyword’ AND table3d.attr49=table12.attr18 AND table3d.attr18=table10c.attr11 AND table3d.attr14=‘keyword’ AND table4d.attr17 IN (‘keyword’) AND table4d.attr19 IN (‘keyword’) AND table16.attr28=table11.attr56 AND table16.attr16=table10b.attr78 AND table16.attr5=table14.attr56 AND table4e.attr34 IN (‘keyword’) AND table4e.attr48 IN (‘keyword’) AND table4f.attr89=table5b.attr7 AND table4f.attr45 IN (‘keyword’) AND table4f.attr1=‘keyword’ AND table10c.attr2=table4e.attr19 AND (table10c.attr78=table12.attr56 OR (table10c.attr55 IS NULL AND table12.attr17 IS NULL)) G. Xiao (FUB) Ontop: Answering SPARQL Queries over Relational Databases 8/56

13. Introduction Overview of Ontop SPARQL Query Answering in Ontop Use Cases Recent Progresses and Future How much time/money is spent searching for data? A user query at Statoil Show all norwegian wellbores with some additional attributes (wellbore id, .....................). Limit to all wellbores with ... and show attributes like ............................................... Limit to all wellbores with ... in ................. and show key attributes in a table. After connecting to ... we could for instance limit further to cores in ... with ...... and where it is larger than a given value, for instance ..... We could also ﬁnd out whether there are cores in ..... which are not stored in .... (based on .....) and where there could be .......... value. Some of the missing data we possibly own, other not. SELECT [...] FROM db_name.table1 table1, db_name.table2 table2a, db_name.table2 table2b, db_name.table3 table3a, db_name.table3 table3b, db_name.table3 table3c, db_name.table3 table3d, db_name.table4 table4a, db_name.table4 table4b, db_name.table4 table4c, db_name.table4 table4d, db_name.table4 table4e, db_name.table4 table4f, db_name.table5 table5a, db_name.table5 table5b, db_name.table6 table6a, db_name.table6 table6b, db_name.table7 table7a, db_name.table7 table7b, db_name.table8 table8, db_name.table9 table9, db_name.table10 table10a, db_name.table10 table10b, db_name.table10 table10c, db_name.table11 table11, db_name.table12 table12, db_name.table13 table13, db_name.table14 table14, db_name.table15 table15, db_name.table16 table16 WHERE [...] table2a.attr1=‘keyword’ AND table3a.attr2=table10c.attr1 AND table3a.attr6=table6a.attr3 AND table3a.attr9=‘keyword’ AND table4a.attr10 IN (‘keyword’) AND table4a.attr1 IN (‘keyword’) AND table5a.kinds=table4a.attr13 AND table5b.kinds=table4c.attr74 AND table5b.name=‘keyword’ AND (table6a.attr19=table10c.attr17 OR (table6a.attr2 IS NULL AND table10c.attr4 IS NULL)) AND table6a.attr14=table5b.attr14 AND table6a.attr2=‘keyword’ AND (table6b.attr14=table10c.attr8 OR (table6b.attr4 IS NULL AND table10c.attr7 IS NULL)) AND table6b.attr19=table5a.attr55 AND table6b.attr2=‘keyword’ AND table7a.attr19=table2b.attr19 AND table7a.attr17=table15.attr19 AND table4b.attr11=‘keyword’ AND table8.attr19=table7a.attr80 AND table8.attr19=table13.attr20 AND table8.attr4=‘keyword’ AND table9.attr10=table16.attr11 AND table3b.attr19=table10c.attr18 AND table3b.attr22=table12.attr63 AND table3b.attr66=‘keyword’ AND table10a.attr54=table7a.attr8 AND table10a.attr70=table10c.attr10 AND table10a.attr16=table4d.attr11 AND table4c.attr99=‘keyword’ AND table4c.attr1=‘keyword’ AND table11.attr10=table5a.attr10 AND table11.attr40=‘keyword’ AND table11.attr50=‘keyword’ AND table2b.attr1=table1.attr8 AND table2b.attr9 IN (‘keyword’) AND table2b.attr2 LIKE ‘keyword’% AND table12.attr9 IN (‘keyword’) AND table7b.attr1=table2a.attr10 AND table3c.attr13=table10c.attr1 AND table3c.attr10=table6b.attr20 AND table3c.attr13=‘keyword’ AND table10b.attr16=table10a.attr7 AND table10b.attr11=table7b.attr8 AND table10b.attr13=table4b.attr89 AND table13.attr1=table2b.attr10 AND table13.attr20=’‘keyword’’ AND table13.attr15=‘keyword’ AND table3d.attr49=table12.attr18 AND table3d.attr18=table10c.attr11 AND table3d.attr14=‘keyword’ AND table4d.attr17 IN (‘keyword’) AND table4d.attr19 IN (‘keyword’) AND table16.attr28=table11.attr56 AND table16.attr16=table10b.attr78 AND table16.attr5=table14.attr56 AND table4e.attr34 IN (‘keyword’) AND table4e.attr48 IN (‘keyword’) AND table4f.attr89=table5b.attr7 AND table4f.attr45 IN (‘keyword’) AND table4f.attr1=‘keyword’ AND table10c.attr2=table4e.attr19 AND (table10c.attr78=table12.attr56 OR (table10c.attr55 IS NULL AND table12.attr17 IS NULL)) At Statoil, it takes up to 4 days to formulate a query in SQL. Statoil loses up to 50.000.000e per year because of this!! G. Xiao (FUB) Ontop: Answering SPARQL Queries over Relational Databases 8/56

14. Introduction Overview of Ontop SPARQL Query Answering in Ontop Use Cases Recent Progresses and Future Challenges Accessing Big Data This is what happens: G. Xiao (FUB) Ontop: Answering SPARQL Queries over Relational Databases 9/56

15. Introduction Overview of Ontop SPARQL Query Answering in Ontop Use Cases Recent Progresses and Future Need for Abstraction We need to facilitate access to Data • by abstracting away from how the data is stored, and • by making use of high level views on the data, so called ontologies. G. Xiao (FUB) Ontop: Answering SPARQL Queries over Relational Databases 10/56

16. Introduction Overview of Ontop SPARQL Query Answering in Ontop Use Cases Recent Progresses and Future Ontology Based Data Access Framework . . . . . . . . . . . . ONTOLOGY = global vocabulary + conceptual view DATA SOURCES external and heterogeneous MAPPINGS how to populate the ontology query result Logical transparency in accessing data: • does not know where and how data is stored; • can only see a conceptual view of data. G. Xiao (FUB) Ontop: Answering SPARQL Queries over Relational Databases 11/56

21. Introduction Overview of Ontop SPARQL Query Answering in Ontop Use Cases Recent Progresses and Future Ontop • Is a platform to query databases through ontologies, relying on semantic technologies. • Compliant with the standards of the W3C. • Supports all major relational DBs (Oracle, DB2, Postgres, MySQL, etc.). • Open-source and released under Apache license. • Development of Ontop: development started 6 years ago already well established: • +200 topics in the mail list • +2300 downloads in last 10 months currently being developed in the context of the EU project Optique G. Xiao (FUB) Ontop: Answering SPARQL Queries over Relational Databases 13/56

22. Introduction Overview of Ontop SPARQL Query Answering in Ontop Use Cases Recent Progresses and Future Architecture of Ontop Ontop SPARQL Query Answering Engine (Quest) OWL-API Sesame Storage And Inference Layer (SAIL) API R2RML API OWL-API (OWL Parser) Sesame API (SPARQL Parser) JDBC Protege Optique Platform Sesame Workbench & SPARQL Endpoint Application Layer API Layer Ontop Core Inputs Relational Databases R2RML Mappings OWL 2 QL Ontologies SPARQL Queries Figure: Architecture of the Ontop system G. Xiao (FUB) Ontop: Answering SPARQL Queries over Relational Databases 14/56

23. Introduction Overview of Ontop SPARQL Query Answering in Ontop Use Cases Recent Progresses and Future Databases Ontop supports standard relational database engines via JDBC. • commercial databases: DB2, Oracle, MS SQL Server • open-source databases: PostgreSQL, MySQL, H2, HSQL • federated databases (e.g., Teiid1 or Exareme2 ) to support multiple data sources (e.g., relational databases, XML, CSV, and Web Services). 1 http://teiid.jboss.org 2 http://www.exareme.org G. Xiao (FUB) Ontop: Answering SPARQL Queries over Relational Databases 15/56

24. Introduction Overview of Ontop SPARQL Query Answering in Ontop Use Cases Recent Progresses and Future Example: Hospital Database Table: tbl patient pid name type stage 1 ’Mary’ false 4 2 ’John’ true 1 types: • false for Non-Small Cell Lung Carcinoma (NSCLC) • true for Small Cell Lung Carcinoma (SCLC),. Stage • NSCLC: 1–6 for stages I, II, III, IIIa, IIIb, and IV, respectively; • SCLC: 1 and 2 for stages Limited and Extensive, respectively. G. Xiao (FUB) Ontop: Answering SPARQL Queries over Relational Databases 16/56

25. Introduction Overview of Ontop SPARQL Query Answering in Ontop Use Cases Recent Progresses and Future Ontology • Ontop uses RDFS and OWL 2 QL as ontology languages. • OWL 2 QL is based on the DL-Lite family of lightweight description logics, which guarantees FO-rewritability Example :NSCLC rdfs:subClassOf :LungCancer . :SCLC rdfs:subClassOf :LungCancer . :LungCancer rdfs:subClassOf :Neoplasm . :hasNeoplasm rdfs:domain :Patient . :hasNeoplasm rdfs:range :Neoplasm . :hasName a owl:DatatypeProperty . :hasStage a owl:ObjectProperty . G. Xiao (FUB) Ontop: Answering SPARQL Queries over Relational Databases 17/56

26. Introduction Overview of Ontop SPARQL Query Answering in Ontop Use Cases Recent Progresses and Future Mappings Ontop supports two mapping languages: • W3C RDB2RDF mapping language R2RML • Ontop native mapping language Example (Mappings in Ontop native mapping language) :db1/{pid} a :Patient . ← SELECT pid FROM tbl patient :db1/neoplasm/{pid} a :NSCLC . ← SELECT pid FROM tbl patient WHERE type = false :db1/neoplasm/{pid} a :SCLC . ← SELECT pid FROM tbl patient WHERE type = true :db1/{pid} :hasName {name} . ← SELECT pid, name FROM tbl patient :db1/{pid} :hasNeoplasm :db1/neoplasm/{pid} . ← SELECT pid FROM tbl patient :db1/neoplasm/{pid} :hasStage :stage-IIIa . ← SELECT pid FROM tbl patient WHERE stage = 4 and type = false G. Xiao (FUB) Ontop: Answering SPARQL Queries over Relational Databases 18/56

27. Introduction Overview of Ontop SPARQL Query Answering in Ontop Use Cases Recent Progresses and Future Queries • Ontop supports essentially all features of SPARQL 1.0 as well as the OWL 2 QL entailment regime of SPARQL 1.1. • Implementation of other features of SPARQL 1.1 (e.g., aggregates, property path queries, negation) is working in progress. The following SPARQL query retrieves all the names of all patients who have a neoplasm (tumor) at stage IIIa. SELECT ?name WHERE { ?p a :Patient ; :hasName ?name ; :hasNeoplasm ?tumor . ?tumor a :Neoplasm ; :hasStage :stage -IIIa . } G. Xiao (FUB) Ontop: Answering SPARQL Queries over Relational Databases 19/56

28. Introduction Overview of Ontop SPARQL Query Answering in Ontop Use Cases Recent Progresses and Future Ontop Core API • The core of Ontop is the SPARQL query answering engine Quest. • We will explain the details in the next section. G. Xiao (FUB) Ontop: Answering SPARQL Queries over Relational Databases 20/56

29. Introduction Overview of Ontop SPARQL Query Answering in Ontop Use Cases Recent Progresses and Future API layer of Ontop System developers can use Ontop as a Java library • OWL API is a reference implementation for creating, manipulating, and serializing OWL ontologies. We extended the OWLReasoner Java interface to support SPARQL query answering. • Sesame is a de-facto standard framework for processing RDF data. Ontop implements the Sesame Storage And Inference Layer (SAIL) API supporting inferencing and querying over relational databases. • Available as Maven artifacts from central repository. G. Xiao (FUB) Ontop: Answering SPARQL Queries over Relational Databases 21/56

30. Introduction Overview of Ontop SPARQL Query Answering in Ontop Use Cases Recent Progresses and Future Application Layer of Ontop • Command line interface • Prot´eg´e plugin • Sesame Workbench and SPARQL Endpoint • Optique Platform • Stardog G. Xiao (FUB) Ontop: Answering SPARQL Queries over Relational Databases 22/56

31. Introduction Overview of Ontop SPARQL Query Answering in Ontop Use Cases Recent Progresses and Future Ontop Protégé plugin The Ontop Protégé plugin provides a graphical interface for: • editing mappings • executing SPARQL queries • checking (in)consistency of the ontology • bootstrapping ontologies and mappings from the database • importing and exporting R2RML mappings • materializing RDF triples, etc. G. Xiao (FUB) Ontop: Answering SPARQL Queries over Relational Databases 23/56

32. Introduction Overview of Ontop SPARQL Query Answering in Ontop Use Cases Recent Progresses and Future Mapping Editor in Prot´eg´e G. Xiao (FUB) Ontop: Answering SPARQL Queries over Relational Databases 24/56

33. Introduction Overview of Ontop SPARQL Query Answering in Ontop Use Cases Recent Progresses and Future SPARQL query answering in Prot´eg´e G. Xiao (FUB) Ontop: Answering SPARQL Queries over Relational Databases 25/56

34. Introduction Overview of Ontop SPARQL Query Answering in Ontop Use Cases Recent Progresses and Future Ontop plugin available from Prot´eg´e plugin repository G. Xiao (FUB) Ontop: Answering SPARQL Queries over Relational Databases 26/56

35. Introduction Overview of Ontop SPARQL Query Answering in Ontop Use Cases Recent Progresses and Future Sesame workbench and SPARQL endpoint • Sesame OpenRDF Workbench is a web application for administrating Sesame repositories. • We extended the Workbench to create and manage Ontop repositories. • Such repositories can then be used as standard HTTP SPARQL endpoints. • Currently Ontop only supports Sesame v2, we are working on supporting v4. G. Xiao (FUB) Ontop: Answering SPARQL Queries over Relational Databases 27/56

36. Introduction Overview of Ontop SPARQL Query Answering in Ontop Use Cases Recent Progresses and Future Sesame workbench and SPARQL endpoint • Sesame OpenRDF Workbench is a web application for administrating Sesame repositories. • We extended the Workbench to create and manage Ontop repositories. • Such repositories can then be used as standard HTTP SPARQL endpoints. • Currently Ontop only supports Sesame v2, we are working on supporting v4. G. Xiao (FUB) Ontop: Answering SPARQL Queries over Relational Databases 27/56

37. Introduction Overview of Ontop SPARQL Query Answering in Ontop Use Cases Recent Progresses and Future Screenshot of the Ontop Sesame Workbench G. Xiao (FUB) Ontop: Answering SPARQL Queries over Relational Databases 28/56

38. Introduction Overview of Ontop SPARQL Query Answering in Ontop Use Cases Recent Progresses and Future Ontop in the Optique Architecture G. Xiao (FUB) Ontop: Answering SPARQL Queries over Relational Databases 29/56

39. Introduction Overview of Ontop SPARQL Query Answering in Ontop Use Cases Recent Progresses and Future Ontop in the Optique Architecture Ontop G. Xiao (FUB) Ontop: Answering SPARQL Queries over Relational Databases 29/56

40. Introduction Overview of Ontop SPARQL Query Answering in Ontop Use Cases Recent Progresses and Future Stardog • Stardog is a commercial triplestore developed by complexible, Inc. • Since version 4 released in November 2015, Stardog has integrated Ontop code to support SPARQL queries over virtual RDF graphs. • The Virtual Graph feature is only available in the enterprise edition G. Xiao (FUB) Ontop: Answering SPARQL Queries over Relational Databases 30/56

42. Introduction Overview of Ontop SPARQL Query Answering in Ontop Use Cases Recent Progresses and Future Conceptual Framework of Query Answering by Query Rewriting ONTOLOGY MAPPINGS DATA SOURCES . . . . . . . . . . . . Ontological Query q Rewritten Query SQLRelational Answer Ontological Answer qresult Rewriting Unfolding Evaluation Result Translation G. Xiao (FUB) Ontop: Answering SPARQL Queries over Relational Databases 32/56

49. Introduction Overview of Ontop SPARQL Query Answering in Ontop Use Cases Recent Progresses and Future Ontop Workflow Ontop ON-LINE OFF-LINE Reasoner Ontology Mapping- Optimiser Mappings DB Integrity Constraints Classified Ontology T-mapping SPARQL Query Query Rewriter SQL query SPARQL to SQL Translator Figure: The Ontop workflow • The off-line stage (start-up time) processes the ontology, mappings, and database integrity constraints. • The on-line stage executes SPARQL queries by rewriting to SQL queries G. Xiao (FUB) Ontop: Answering SPARQL Queries over Relational Databases 33/56

50. Introduction Overview of Ontop SPARQL Query Answering in Ontop Use Cases Recent Progresses and Future Offline Stage The offline stage can be thought of as consisting of three phases: • ontology classification • T-mapping construction • T-mapping optimization Example • New axioms in the classified Ontology :NSCLC rdfs:subClassOf :Neoplasm . :SCLC rdfs:subClassOf :Neoplasm . • Inferred Mappings after T-mapping construction :db1/neoplasm/{pid} a :Neoplasm . ← SELECT pid FROM tbl patient WHERE type = false :db1/neoplasm/{pid} a :Neoplasm . ← SELECT pid FROM tbl patient WHERE type = true • Optimized T-mappings :db1/neoplasm/{pid} a :Neoplasm . ← SELECT pid FROM tbl patient WHERE type = false OR type = true G. Xiao (FUB) Ontop: Answering SPARQL Queries over Relational Databases 34/56

51. Introduction Overview of Ontop SPARQL Query Answering in Ontop Use Cases Recent Progresses and Future Online Stage During query execution (the online stage), Ontop transforms an input SPARQL queries into an optimized SQL query using the T-mappings and database integrity constraints. Optimizing the generated SQL queries structural optimizations • pushing the joins inside the unions, • pushing the functions as high as possible in the query tree, • eliminating sub-queries. Semantic query optimizations semantic analysis of SQL queries to reduce the size and complexity • removing redundant self-joins, • detecting unsatisﬁable or trivially valid (true) conditions. G. Xiao (FUB) Ontop: Answering SPARQL Queries over Relational Databases 35/56

52. Introduction Overview of Ontop SPARQL Query Answering in Ontop Use Cases Recent Progresses and Future Example of SQL translation and optimization • Consider a SPARQL query SELECT ?x WHERE { ?x a :Neoplasm ; :hasStage :stage -IIIa . } • Non-optimized generated SQL query SELECT Q1.x FROM (( SELECT concat(":db1/neoplasm/", pid) AS x FROM tbl_patient WHERE type = false OR type = true) Q1 JOIN (SELECT concat(":db1/neoplasm/", pid) AS x FROM tbl_patient WHERE stage = 4 AND type = false) Q2 ON Q1.x = Q2.x) • SQL query after the structural optimization SELECT concat(":db1/neoplasm/", Q.pid) AS x FROM (SELECT T1.pid FROM tbl_patient T1 JOIN tbl_patient T2 ON T1.pid = T2.pid WHERE (T1.type = false OR T1.type = true) AND T2.stage = 4 AND T2.type = false) Q • SQL query after the self-join elimination SELECT concat(":db1/neoplasm/", Q.pid) AS x FROM (SELECT pid FROM tbl_patient WHERE type = false AND stage = 4) Q • SQL query after the second structural optimization SELECT concat(":db1/neoplasm/", pid) AS x FROM tbl_patient WHERE type = false AND stage = 4 G. Xiao (FUB) Ontop: Answering SPARQL Queries over Relational Databases 36/56

58. Introduction Overview of Ontop SPARQL Query Answering in Ontop Use Cases Recent Progresses and Future Statoil Use Case • Optique Use Case Partner • Main reference: “Ontology Based Access to Exploration Data at Statoil [Kharlamov, Hovland, et al., 2015, ISWC In-use Track]. • Exploration domain • Improve the eﬃciency of the information gathering routine for geologists at Statoil • Eﬃcient, creative data collection from multiple data sources • ⇒ separate slides for this use case G. Xiao (FUB) Ontop: Answering SPARQL Queries over Relational Databases 38/56

59. Introduction Overview of Ontop SPARQL Query Answering in Ontop Use Cases Recent Progresses and Future Siemens Use Case • Optique Use Case Partner • Main reference: “How Semantic Technologies Can Enhance Data Access at Siemens Energy” [Kharlamov, Solomakhina, et al., 2014, ISWC In-use Track] • ⇒ separate slides for this use case G. Xiao (FUB) Ontop: Answering SPARQL Queries over Relational Databases 39/56

60. Introduction Overview of Ontop SPARQL Query Answering in Ontop Use Cases Recent Progresses and Future EPNet Use Case • EPNet Project (ERC Advanced Grant EPNet “Production and distribution of food during the Roman Empire: Economics and Political Dynamics”, ERC-2013-ADG 340828). • Main reference: “Ontology-Based Data Integration in EPNet: Production and Distribution of Food During the Roman Empire” [Calvanese, Liuzzo, et al., 2016, J. of Eng. Appl. of AI] • Ontology-Based Data Integration: integrating multiple datasource. • Linking three datasets: the EPNet relational repository , the Epigraphic Database Heidelberg, and the Pleiades dataset G. Xiao (FUB) Ontop: Answering SPARQL Queries over Relational Databases 40/56

61. Introduction Overview of Ontop SPARQL Query Answering in Ontop Use Cases Recent Progresses and Future EMSec Use Case • EMSec (Echtzeitdienste f¨ur die Maritime Sicherheit, Real-time Services for the Maritime Security) is a German BMBF (Federal Ministry of Research and Education) funded project • Geo-spatial support by Ontop-spatial (developed as a fork of Ontop) • Sextant for visualizing linked geospatial data • Use case paper “Ontology-based Data Access for Maritime Security” is under submission G. Xiao (FUB) Ontop: Answering SPARQL Queries over Relational Databases 41/56

62. Introduction Overview of Ontop SPARQL Query Answering in Ontop Use Cases Recent Progresses and Future IBM Research Ireland Use Case • Main reference: “Data Access Linking and Integration with DALI: building a Safety Net for an Ocean of City Data” [Lopez et al., 2015, ISWC In-use Track] • Smarter Cities Technology Centre, IBM Research, Ireland G. Xiao (FUB) Ontop: Answering SPARQL Queries over Relational Databases 42/56

63. Introduction Overview of Ontop SPARQL Query Answering in Ontop Use Cases Recent Progresses and Future Electronic Health Records Use Case • Main reference: “Validating an ontology-based algorithm to identify patients with Type 2 Diabetes Mellitus in Electronic Health Records” [Rahimi et al., 2014, Int. J. of Medical Informatics] • Medicine, The University of New South Wales, Australia G. Xiao (FUB) Ontop: Answering SPARQL Queries over Relational Databases 43/56

64. Introduction Overview of Ontop SPARQL Query Answering in Ontop Use Cases Recent Progresses and Future Electronic Health Records Use Case (Cont.) G. Xiao (FUB) Ontop: Answering SPARQL Queries over Relational Databases 44/56

65. Introduction Overview of Ontop SPARQL Query Answering in Ontop Use Cases Recent Progresses and Future Use Cases • More use cases are in https://github.com/ontop/ontop/wiki/UseCases • Unfortunately, we are not able to track all use cases of Ontop. G. Xiao (FUB) Ontop: Answering SPARQL Queries over Relational Databases 45/56

67. Introduction Overview of Ontop SPARQL Query Answering in Ontop Use Cases Recent Progresses and Future Recent Progresses More recent lines of research on Ontop include • formalization of SPARQL in the context of OBDA [Rodriguez-Muro and Rezk, 2015, J. Web Semantics] [Kontchakov et al., 2014, ISWC] • OWL 2 QL entailment regime [Kontchakov et al., 2014, ISWC] • SWRL rule language with a limited form of recursion handled by SQL Common Table Expressions [Xiao et al., 2014, RR] • owl:sameAs for cross-linked datasets [Calvanese, Giese, et al., 2015, ISWC] • Expressive ontologies beyond OWL 2 QL by rewriting and approximation with the help of the mapping layer [Botoeva et al., 2016, AAAI] • System description of Ontop [Calvanese, Cogrel, et al., 2016, Semantic Web J., to appear] G. Xiao (FUB) Ontop: Answering SPARQL Queries over Relational Databases 47/56

68. Introduction Overview of Ontop SPARQL Query Answering in Ontop Use Cases Recent Progresses and Future Recent Progresses More recent lines of research on Ontop include • formalization of SPARQL in the context of OBDA [Rodriguez-Muro and Rezk, 2015, J. Web Semantics] [Kontchakov et al., 2014, ISWC] • OWL 2 QL entailment regime [Kontchakov et al., 2014, ISWC] • SWRL rule language with a limited form of recursion handled by SQL Common Table Expressions [Xiao et al., 2014, RR] • owl:sameAs for cross-linked datasets [Calvanese, Giese, et al., 2015, ISWC] • Expressive ontologies beyond OWL 2 QL by rewriting and approximation with the help of the mapping layer [Botoeva et al., 2016, AAAI] • System description of Ontop [Calvanese, Cogrel, et al., 2016, Semantic Web J., to appear] G. Xiao (FUB) Ontop: Answering SPARQL Queries over Relational Databases 47/56

69. Introduction Overview of Ontop SPARQL Query Answering in Ontop Use Cases Recent Progresses and Future Beyond OWL 2 QL (AAAI 16 paper) • Framework for Rewriting and Approximation of OBDA specifications . . . . . . . . . . . . T , M, S T , M , S ⇒ . . . . . . . . . . . . Rewriting The new specification is equivalent to the original one w.r.t. query answering (query-inseparable). Approximation The new specification is a sound approximation of the original one w.r.t. query answering. G. Xiao (FUB) Ontop: Answering SPARQL Queries over Relational Databases 48/56

70. Introduction Overview of Ontop SPARQL Query Answering in Ontop Use Cases Recent Progresses and Future Beyond OWL 2 QL (II) G. Xiao (FUB) Ontop: Answering SPARQL Queries over Relational Databases 49/56

71. Introduction Overview of Ontop SPARQL Query Answering in Ontop Use Cases Recent Progresses and Future Beyond OWL 2 QL (III) G. Xiao (FUB) Ontop: Answering SPARQL Queries over Relational Databases 50/56

72. Introduction Overview of Ontop SPARQL Query Answering in Ontop Use Cases Recent Progresses and Future WIP: OBDA beyond Relational Databases Mapping parsing a Ontology parsing b Mapping compilation c sparql parsing 0 Rewriting 1 Unfolding w.r.t. mappings 2 Structural/semantic optimization 3 Normalization/ Decomposition 4 RA-to-native query translation 5 Evaluation 6 Post- processing 7 Mapping ﬁle Ontology ﬁle Mapping M Ontology T T -Mapping MT sparql string sparql Q Rewritten sparql QT RA q1 RA q2 RA q3Native queries Native results sparql result OFFLINE ONLINE • NoSQL Movement • In fact most of the components of Ontop are SQL-independent • We are working on OBDA over non-relational datasource • We are targeting on MongoDB now G. Xiao (FUB) Ontop: Answering SPARQL Queries over Relational Databases 51/56

73. Introduction Overview of Ontop SPARQL Query Answering in Ontop Use Cases Recent Progresses and Future Future • In order to further improve performance, we will investigate data-dependent optimizations. • support larger fragments of SPARQL (e.g., aggregation, negation, and path queries) and R2RML (e.g., named graphs). • For end-users, we will improve the GUI and extend utilities to make Ontop even more user-friendly. • go beyond relational databases and support other kinds of data sources (e.g., graph and document databases). • Continue building community G. Xiao (FUB) Ontop: Answering SPARQL Queries over Relational Databases 52/56

74. Acknowledgment

75. Thank you for your attention! QUESTIONS?

76. Introduction Overview of Ontop SPARQL Query Answering in Ontop Use Cases Recent Progresses and Future References I Kharlamov, Evgeny, Nina Solomakhina, Özgür Lütfü Öz¸cep, Dmitriy Zheleznyakov, Thomas Hubauer, Steffen Lamparter, Mikhail Roshchin, Ahmet Soylu, and Stuart Watson (2014). “How Semantic Technologies Can Enhance Data Access at Siemens Energy”. In: The Semantic Web - ISWC 2014 - 13th International Semantic Web Conference, Riva del Garda, Italy, October 19-23, 2014. Proceedings, Part I, pp. 601–619. Kontchakov, Roman, Martin Rezk, Mariano Rodriguez-Muro, Guohui Xiao, and Michael Zakharyaschev (2014). “Answering SPARQL Queries over Databases under OWL 2 QL Entailment Regime”. In: vol. 8796. doi:10.1007/978-3-319-11964-9 35, pp. 552–567. Rahimi, Alireza, Siaw-Teng Liaw, Jane Taggart, Pradeep Ray, and Hairong Yu (2014). “Validating an ontology-based algorithm to identify patients with Type 2 Diabetes Mellitus in Electronic Health Records”. In: Int. J. of Medical Informatics 83.10. doi:10.1016/j.ijmedinf.2014.06.002, pp. 768–778. Xiao, Guohui, Martin Rezk, Mariano Rodriguez-Muro, and Diego Calvanese (2014). “Rules and Ontology Based Data Access”. In: Proc. 8th Int. Conference on Web Reasoning and Rule Systems (RR 2014). Ed. by Marie-Laure Mugnier and Roman Kontchakov. Lecture Notes in Computer Science. Springer. Calvanese, Diego, Martin Giese, Dag Hovland, and Martin Rezk (2015). “Ontology-based Integration of Cross-linked Datasets”. In: Proc. of the 14th Int. Semantic Web Conference (ISWC). Lecture Notes in Computer Science. Springer. Kharlamov, Evgeny, Dag Hovland, et al. (2015). “Ontology Based Access to Exploration Data at Statoil”. In: The Semantic Web - ISWC 2015 - 14th International Semantic Web Conference, Bethlehem, PA, USA, October 11-15, 2015, Proceedings, Part II, pp. 93–112. Lopez, Vanessa, Martin Stephenson, Spyros Kotoulas, and Pierpaolo Tommasi (2015). “Data Access Linking and Integration with DALI: Building a Safety Net for an Ocean of City Data”. In: The Semantic Web - ISWC 2015 - 14th International Semantic Web Conference, Bethlehem, PA, USA, October 11-15, 2015, Proceedings, Part II, pp. 186–202. Rodriguez-Muro, Mariano and Martin Rezk (2015). “Efficient SPARQL-to-SQL with R2RML Mappings”. In: 33. doi:10.1016/j.websem.2015.03.001, pp. 141–169. G. Xiao (FUB) Ontop: Answering SPARQL Queries over Relational Databases 55/56

77. Introduction Overview of Ontop SPARQL Query Answering in Ontop Use Cases Recent Progresses and Future References II Botoeva, Elena, Diego Calvanese, Valerio Santarelli, Domenico Fabio Savo, Alessandro Solimando, and Guohui Xiao (2016). “Beyond OWL 2 QL in OBDA: Rewritings and Approximations”. In: Proc. of the 30th AAAI Conf. on Artiﬁcial Intelligence (AAAI). AAAI Press. Calvanese, Diego, Benjamin Cogrel, Sarah Komla-Ebri, Roman Kontchakov, Davide Lanti, Martin Rezk, Mariano Rodriguez-Muro, and Guohui XIao (2016). “Ontop: Answering SPARQL Queries over Relational Databases”. In: Semantic Web Journal. Calvanese, Diego, Pietro Liuzzo, Alessandro Mosca, Jose Remesal, Martin Rezk, and Guillem Rull (2016). “Ontology-Based Data Integration in EPNet: Production and Distribution of Food During the Roman Empire”. In: Engineering Applications of Artiﬁcial Intelligence. G. Xiao (FUB) Ontop: Answering SPARQL Queries over Relational Databases 56/56