Linking application and semantic data with RDF Lens

Abstract

Linked Data is commonly regarded as an unfriendly data structure to be directly used by application developers. The (often unknown) triple-based structure of RDF graphs causes developers to struggle to extract the triples of interest and translate them into the object-like structure needed for their application. A generic, composable and reusable way to look into RDF graphs as plain objects would remove an important barrier for integrating Linked Data in all facets of application logic. We propose RDF Lens, a library based on ideas from the Haskell lens library that allows for composable and reusable data extraction units, called lenses. Value is shown by implementing a lens that generates a new lens based on a SHACL shape that extracts the semantic data into the desired plain object. Abstracting data extraction at the lens level allows for mixed extraction: using both custom extraction

and declarative extraction, which could increase ease of use and reusability. The current implementation is a proof of concept that defines how to extract data from an RDF graph in JavaScript applications but does not allow (yet) writing or altering linked data with the same lenses. Future work would allow for creating and updating Linked Data in the same elegant way