Client-side Processing of GeoSPARQL Functions with Triple Pattern Fragments
2 likes801 views
The document discusses the limitations and potential enhancements of client-side geospatial processing using Triple Pattern Fragments (TPF) to support GeoSPARQL functions, highlighting the collaborative project between ADAPT and Ordnance Survey Ireland. While existing implementations of TPF do not inherently support GeoSPARQL, the authors explore two approaches for extending TPF to accommodate these functions, emphasizing the need for a pure JavaScript implementation to facilitate stakeholder access. Initial findings suggest that while their modified approach is viable, further investigation into server-side support for GeoSPARQL and performance benchmarking is required.
Client-side Processing of GeoSPARQL Functions with Triple Pattern Fragments
- 2. www.adaptcentre.ieIntroduc9on Geospa9al informa9on is an important part of the LD Web Some geographic LD datasets are authorita9ve • Ordnance Survey Linked Data (UK) , data.geohive.ie (IE) Not all geographic LD datasets adopt standards • Ordnance Survey Linked Data uses their own vocabulary, and provides no support for spaXal funcXons Not all endpoints serving GeoSPARQL support its func9ons • GeospaXal funcXons are computaXonal expensive; server load • One o[en provides data dumps and resolvable URIs as a “good enough” pracXce to avoid this problem (Verborgh et al. 2016) R. Verborgh, M. Vander Sande, O. HarXg, J. Van Herwegen, L. De Vocht, B. De Meester, G. Haesendonck, and P. Colpaert. 2016. Triple Pacern Fragments: A low-cost knowledge graph interface for the Web. J. Web Sem. 37-38 (2016), 184–206
- 4. www.adaptcentre.ieIntroduc9on – data.geohive.ie data.geohive.ie • Provides resolvable URIs, dumps and a LD frontend • Decided against providing a public SPARQL endpoint • Triple Pacern Fragments were adopted instead Triple Pa?ern Fragments (TPF) • Distribute load between a TPF client and server • Less load on the server at the cost of increased bandwidth
- 6. www.adaptcentre.ieApproach Two possible approaches: A) Extending a TPF Client • TPF server specificaXon intact (backwards compaXble) • Possibly more network overhead B) Extending the TPF server • Outside server specificaXon, but proven to be viable for substring filtering (Van Herwegen et al. 2015) Addi9onal requirement: a pure JavaScript implementa9on • Allows one to run the client in a browser and hence facilitate stakeholders in formulaXng GeoSPARQL queries J. Van Herwegen, L. De Vocht, R. Verborgh, E. Mannens, and R. Van de Walle. 2015. Substring Filtering for Low-Cost Linked Data Interfaces. In The SemanXc Web - ISWC 2015 - 14th InternaXonal SemanXc Web Conference, Bethlehem, PA, USA, October 11-15, 2015, Proceedings, Part I (LNCS), Vol. 9366. Springer, 128–143.
- 7. www.adaptcentre.ieApproach Implementa9on • We extended V2.0.4 of the TPF Node.js Client (Available at hcps://github.com/chrdebru/Client.js) • Made available a web-client using this extension (At hcp://theme-e.adaptcentre.ie/geo-tpf/) GeoSPARQL coverage • Support for GeoSPARQL FILTER funcXons that we deem to occur o[en in examples (sfTouches, sfWithin, …) Func9ons are implemented in terms of set opera9ons • E.g., sfOverlaps: The intersecXon of the 2 geometries is not empty and should contain polygons or mulX-polygons denoXng areas.
- 9. www.adaptcentre.ieDiscussion (condensed) On performance • Based on iniXal experiments. Most Xme went to processing of GeoSPARQL FILTER funcXons. Slightly worse results in terms of Xme compared to running queries in a triplestore. We thus deem our approach viable. • However, our approach using JavaScript (in a browser) will unlikely outperform such triplestores in terms of performance On Op9miza9on • OSi provides high-res polygons, which causes a lot of network traffic. • br-TPF (HarXg and Buil Aranda, 2016) did reduce the number of HTTP calls, but not really the amount of traffic for the polygons. • Server-side support for GeoSPARQL funcXons could tackle that problem O. HarXg and C. Buil Aranda. 2016. Bindings-Restricted Triple Pacern Fragments. In On the Move to Meaningful Internet Systems: OTM 2016 Conferences - Confederated InternaXonal Conferences: CoopIS, C&TC, and ODBASE 2016, Rhodes, Greece, October 24-28, 2016, Proceedings (LNCS), Vol. 10033. Springer, 762–779.
- 10. www.adaptcentre.ieConclusions and Future Work Conclusions • We extended a TPF client to support GeoSPARQL funcXons • The paper illustrates how we used it for different projects Lessons Learned • IniXal results showed our approach is viable • Though network overhead may sXll be an issue, an extension of the TPF server is worthwhile exploring (future work) • How to deal with GeoSPARQL predicates and funcXons? Future work • Complete implementaXon and test with funcXonal benchmarks, e.g., Geographica (Garbis et al, 2013) • Experiments involving stakeholders, and idenXfy the limitaXons of our approach to characterize, for instance, “sensible” queries G. Garbis, K. Kyzirakos, and M. Koubarakis. 2013. Geographica: A Benchmark for GeospaXal RDF Stores (Long Version). In the SemanXc Web - ISWC 2013 - 12th InternaXonal SemanXc Web Conference, Sydney, NSW, Australia, October 21-25, 2013, Proceedings, Part II (LNCS), Vol. 8219. Springer, 343–359.