The Highway to Queryable Linked Data: Self-Describing Web APIs with Varying Features

Making Linked Data queryable on the Web is not an easy task for publishers, for technical and logistical reasons. Can they afford to offer a sparql endpoint, or should they offer an api or data dump instead? And what technical knowledge is needed for that? This demo presents a user-friendly pipeline to compose apis for Linked Datasets, consisting of a customizable set of reusable features, e.g., Triple Pattern Fragments, substring search, membership metadata, etc. These apis indicate their supported features in hypermedia responses, so that clients can discover which server-provided functionality they understand, and divide the evaluation of sparql queries accordingly between client and server. That way, publishers can determine the complexity of the resulting api, and thus the maximal set of server tasks. This demo shows how publishers can easily set up an api with this pipeline, and demonstrates the client-side execution of federated sparql queries against such apis.