Abstract :
[en] KG construction often involves mapping data from various sources to RDF, which can be achieved with declarative languages such as RML.
RML, while powerful, presents authoring difficulties due to its RDF graph-based structure and Turtle serialization.
Tools and languages have been proposed to help the creation of mappings by creating (UI) abstraction of RML concepts or more accessible languages.
However, existing tools often struggle to balance guidance and flexibility.
While representing RML in RDF allows enriching mappings with other vocabularies, it also presents a steep learning curve for users unfamiliar with Turtle and RML.
This paper introduces GRAPE, an open-source projectional editor designed to facilitate RML mapping creation while preserving the flexibility of RDF schemas used to extend RML.
GRAPE uses a language-oriented approach, treating each RML module as a domain-specific language.
This allows users to write RML and/or Turtle in one artifact.
Implemented using JetBrains MPS, GRAPE provides a text-like projection for manipulating the abstract syntax tree.
GRAPE ensures syntactic correctness, prevents errors and guides with auto-completion and intentions.
This lowers the barrier for new users while preserving the flexibility required for complex mappings.
This paper aims to demonstrate the feasibility of this approach.
It also describes the steps we have conducted to create an editor for Turtle, which is then extended for RML and its modules.
The approach is extensible and allows the community to develop languages for other vocabularies.
Scopus citations®
without self-citations
0
