Low-code application platform¶

LinkedDataHub is also a low-code platform for serving RDF Knowledge Graph dataspaces and building applications on top of them. It provides development and management features:

• rich UI component library that can be customized without writing code
• SPARQL-based templates for instance creation and editing forms (constructors)
• SPIN and SHACL constraints for validation of incoming data
• authentication using WebID and OpenID Connect (Google and ORCID)
• access control using WAC authorizations
• declarative packages that plug vocabulary support and custom UI into a dataspace with a single triple
• Git-backed document versioning with the standard Memento protocol for accessing the history
• the ldh command line interface that drives the same HTTP API as the UI

Knowledge management and collaboration¶

As an RDF-native CMS, LinkedDataHub provides these end-user features:

• creating structured content from Linked Data and SPARQL resources
• creating RDF documents (graphs)
• browsing Linked Data
• copying RDF resources into the local dataspace
• searching and querying SPARQL endpoints
• creating collections from SPARQL queries
• editing content in place, with RDFa annotations linking text to resources
• importing CSV data by mapping it to RDF using SPARQL
• visualizing query results as graphs, charts etc.
• viewing, diffing and restoring document versions

Architecture¶

Architecturally, LinkedDataHub is a read-write RDF Graph Store combined with a rich Linked Data/SPARQL client. LinkedDataHub does not persist RDF data itself but rather serves it from, and stores it in, a backing triplestore, which by default is Apache Jena Fuseki.

Every document in a LinkedDataHub dataspace is also a named graph in the Graph Store and has both RDF and HTML representations.

The client is implemented using XSLT 3.0, a standard, declarative data transformation language. The same stylesheets run in two environments: on the server (Saxon-HE) to produce the initial HTML response, and in the browser (Saxon-JS 3 with Interactive XSLT) to render the layout and drive all subsequent navigation without full page reloads. The client can connect to any Linked Data resource or SPARQL 1.1 endpoint.

LinkedDataHub no longer uses Linked Data Templates (support was removed in version 3.0.3). However, they can still be used to publish Linked Data from SPARQL endpoints using Processor.

See the CHANGELOG for release notes.