Designing scientific SPARQL queries using autocompletion by snippets

Abstract : SPARQL is the standard query language used to access RDF linked data sets available on the Web. However, designing a SPARQL query can be a tedious task, even for experienced users. This is often due to imperfect knowledge by the user of the ontologies involved in the query. To overcome this problem, a growing number of query editors offer autocompletion features. Such features are nevertheless limited and mostly focused on typo checking. In this context, our contribution is four-fold. First, we analyze several autocompletion features proposed by the main editors, highlighting the needs currently not taken into account while met by a user community we work with, scientists. Second, we introduce the first (to our knowledge) autocompletion approach able to consider snippets (fragments of SPARQL query) based on queries expressed by previous users, enriching the user experience. Third, we introduce a usable, open and concrete solution able to consider a large panel of SPARQL autocompletion features that we have implemented in an editor. Last but not least, we demonstrate the interest of our approach on real biomedical queries involving services offered by the Wikidata collaborative knowledge base.
Type de document :
Communication dans un congrès
14th IEEE International Conference on eScience, Oct 2018, Amsterdam, Netherlands. 2018
Liste complète des métadonnées

https://hal.archives-ouvertes.fr/hal-01874780
Contributeur : Karima Rafes <>
Soumis le : samedi 15 septembre 2018 - 08:26:23
Dernière modification le : lundi 4 mars 2019 - 11:36:08
Document(s) archivé(s) le : dimanche 16 décembre 2018 - 12:23:18

Fichier

papier-long-e (18).pdf
Fichiers produits par l'(les) auteur(s)

Identifiants

  • HAL Id : hal-01874780, version 2

Citation

Karima Rafes, Serge Abiteboul, Sarah Cohen-Boulakia, Bastien Rance. Designing scientific SPARQL queries using autocompletion by snippets. 14th IEEE International Conference on eScience, Oct 2018, Amsterdam, Netherlands. 2018. 〈hal-01874780v2〉

Partager

Métriques

Consultations de la notice

397

Téléchargements de fichiers

104