R4R: Template-based REST API Framework for RDF Knowledge Graphs

Abstract

Knowledge graphs (KGs) are increasingly being used to make structured information available on the Web, by means of REST APIs and/or SPARQL endpoints. In many cases, these REST APIs are generated on top of the SPARQL endpoints, using existing technology approaches that are based on proprietary con guration les or ontologies to create the APIs. These approaches may impose content-based or structural constraints when composing Web resources. To relax these constraints we propose R4R, a more exible solution based on Web standards and REST principles that creates and publishes customizable APIs exposing Web resources from SPARQL queries organized in le system directories. R4R features include individual and nested resources, paginated queries, optional elds, web authentication, query parameters and sorting lists. Resource type: Software License: Apache License 2.0 DOI: https://doi.org/10.5281/zenodo.3543320

Keywords

• API
• Knowledge Graph
• REST
• SPARQL

-

Knowledge graphs (KGs) are drawing increasing attention from both academia and industry for representing, sharing and using knowledge in applications [ 7, 3 ]. They may be made available as RDF-based datasets, including a SPARQL endpoint (e.g., DBpedia), and/or via REST APIs (e.g., Google Knowledge Graph). In both cases, KGs share many commonalities from the data representation point of view (both use triples to represent facts), but they are radically di erent in terms of query capabilities: SPARQL provides a more expressive query language than what can be normally done with a REST API, but it can be a barrier for non-expert users. Web APIs usually present data according to REpresentative State Transfer (REST) architecture principles mapping HTTP verbs (POST, GET, PUT, DELETE) to CRUD operations (Create, Read, Update, Delete).

Copyright © 2021 for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0) However, API resources do not necessarily match to KG resources. A public procurement-focused API and a technology-focused API may present, in a different way, the information retrieved from the same KG about companies. One more focused on merit and the other on innovations.

In this demo we present R4R, an open source framework that facilitates the publication of a KG via a REST API over HTTP. Our approach proposes a fully customizable de nition of resources, both naming and content (and even nesting), through a hierarchical organization of data (Figure 1). It deploys a web service based on SPARQL queries to retrieve the information and provides templates to compose resources that are organized in folders in a system directory. Finally, we describe a motivating example where R4R is used to enhance KG access. 2

KG data consumption via Web API Several approaches are available to provide Web developers with mechanisms to ease KG data consumption without dealing with the complexity of Semantic Web standards and technologies, namely SPARQL. Some of these approaches have been focused on the provision of Web APIs that allow developers to interact with KG data. There are tools [ 1 ][ 5 ] that generate Web APIs from set of SPARQL queries but require setting up programming environments or use xed structures for resources. Others [ 4 ] proposed a SPARQL Transformer to provide speci c JSON structures from the SPARQL queries but this prevents validating SPARQL queries from any SPARQL endpoint, or [ 2 ] creates a Web service from the API paths, methods and SPARQL queries provided in a con guration le based on a key-value structure in proprietary format. In Table 1 we summarized the main features of the aforementioned approaches, and it also includes the features of R4R which will be described in the following sections.

Characteristic

Resources

Methods

Security Documentation

Parameters

Serialization

Content

Deployment

3 R4R

Proposal

single multiple nested GET POST

PUT

DELETE basic authentication Swagger-compliant static HTML

ltering ordering pagination

XML CSV JSON

RDF

JSON-oriented SPARQL compatible remote local isolated

X X x X X x x X X x x x x X X X X x x X X x

X X x X X x x X X x x x X x X X X X x X X X

X X x X X x x x x X X X x x X X x X x X X x

X X X X x x x X X X X X X x x X x X X X X X The R4R1 open source framework aims to bring it closer to non-expert Web service developers: (1) customizable resource abstraction and an (2) intuitive REST-based interface. The purpose of this tool is to facilitate access to KGs guided by use cases via a REST API. R4R supports isolated deployment without the need to con gure programming environments. Unlike existing approaches, R4R provides nested resources that let us reference complex objects. Thanks to this feature, users can, for example, get the characters of a movie through an API call like /movies/fidg/characters where the resulting characters will depend on the particular movie resource identi ed with the \fidg" value. It also maintains SPARQL compatibility by avoiding any non-SPARQL variables in queries, so that queries can be externally validated from any SPARQL endpoint. Our approach allows users to generate API documentation from a YAML le with a Swagger speci cation as well as with a static HTML le, for those users not familiar with the Swagger speci cation. In regard to parameters, our approach provides ltering, ordering, and pagination options which allows full exibility when users have to deal with resources. 4

Motivating Example We have prepared a short tutorial1 to create a REST API over DBpedia that browses movies. Following an intuitive approach to the REST architectural style, 1 https://github.com/oeg-upm/r4r the characters of a movie, for example, are available thanks to a SPARQLquery (Listing 1.1) and a template (Listing 1.2) les located in a resources/movies/characters folder. As a result, a JSON message with the list of characters of the movie is obtained by requesting, for example, /movies/WarGames/characters (Listing 1.3).

Listing 1.1. SPARQL query to retrieve characters of a movie } , } , { } , { } PREFIX dbo: <http://dbpedia.org/ontology/> PREFIX dbp: <http://dbpedia.org/property/> PREFIX res: <http://dbpedia.org/resource/> PREFIX dbr: <http://dbpedia.org/resource/> PREFIX rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> PREFIX foaf: <http://xmlns.com/foaf/0.1/> PREFIX rdfs: <http://www.w3.org/2000/01/rdf-schema#> SELECT ?name ?birthDate ( ?starring AS ?uri ) WHERE { ?id dbo:starring ?starring . ?starring foaf:name ?name . ?starring dbo:birthDate ?birthDate .

OPTIONAL {?name rdfs:label ?string . FILTER (lang(?string) = 'en') }

Listing 1.2. Template to return characters of a movie #foreach( $person in $results ) { "uri" : "$person.uri", "name" : "$person.name", "birthDate" : "$person.birthDate" } #if ( $velocityCount < ${results.size()} )

, #end #end

Listing 1.3. Main characters of the movie WarGames "uri" : "http://dbpedia.org/resource/John_Wood_(English_actor)", "name" : "John Wood", "birthDate" : "1930-07-05"

{ "uri" : "http://dbpedia.org/resource/Ally_Sheedy", "name" : "Ally Sheedy", "birthDate" : "1962-06-13" "uri" : "http://dbpedia.org/resource/Matthew_Broderick", "name" : "Matthew Broderick", "birthDate" : "1962-03-21" "uri" : "http://dbpedia.org/resource/Dabney_Coleman", "name" : "Dabney Coleman", "birthDate" : "1932-01-03" R4R facilitates the creation and publication of Web resources following REST principles through lesystem directories. Resource paths are constrained to follow a tree structure by automatically creating them from directories and avoiding manual editing. In addition, R4R eases KG consumption by providing a developer-friendly serialization of the SPARQL results without imposing custom con guration rules but using standard SPARQL and Velocity data structures. It has been adopted by TheyBuyForYou (TBFY) project [ 6 ] to facilitate access to its public procurement KG2 with more than 150 million triples.

R4R has to be seen as a fully operational rst step of building REST API over KG based on templates to create resources, SPARQL queries to retrieve data and directories to de ne resource paths. We plan to extend R4R in the future to enable write operations (POST, PUT, DELETE) to fully interact with the KG data and implement additional content negotiation capabilities and formats (JSON-LD, Turtle, HTML).

Acknowledgments Work supported by KnowledgeSpaces, PID2020-118274RB-I00. 2 https://github.com/TBFY/knowledge-graph-API

References

1. Daga , E. , Panziera , L. , Pedrinaci , C. : A BASILar approach for building web APIs on top of SPARQL endpoints . In: CEUR Workshop Proceedings . vol. 1359 , pp. 22 { 32 ( 2015 )

2. Daquino , M. , Heibi , I. , Peroni , S. , Shotton , D. : Creating Restful APIs over SPARQL endpoints with RAMOSE . arXiv preprint arXiv: 2007 . 16079 ( 2020 )

3. Hogan , A. , Blomqvist , E. , Cochez , M., d'Amato , C. , de Melo , G., Gutierrez , C. , Gayo , J.E.L. , Kirrane , S. , Neumaier , S. , Polleres , A. , Navigli , R. , Ngomo , A.C.N. , Rashid , S.M. , Rula , A. , Schmelzeisen , L. , Sequeda , J. , Staab , S. , Zimmermann , A. : Knowledge Graphs ( 2021 )

4. Lisena , P. , Meron~ o-Pen~uela, A. , Kuhn , T. , Troncy , R.: Easy Web API Development with SPARQL Transformer . In: Ghidini, C. , Hartig , O. , Maleshkova , M. , Svatek , V. , Cruz , I. , Hogan , A. , Song , J. , Lefrancois , M. , Gandon , F . (eds.) The Semantic Web { ISWC 2019 . pp. 454 { 470 . Springer International Publishing, Cham ( 2019 )

5. Meron ~o-Pen~uela, A. , Hoekstra , R.: grlc makes GitHub taste like linked data APIs . In: European Semantic Web Conference . pp. 342 { 353 . Springer ( 2016 )

6. Soylu , A. , Corcho , O. , Elves ter , B. , Badenes-Olmedo , C. , Mart nez , F. Y. , Kovacic , M. , Posinkovic , M. , Makgill , I. , Taggart , C. , Simperl , E. , Lech , T.C. , Roman , D. : Enhancing public procurement in the european union through constructing and exploiting an integrated knowledge graph . In: The Semantic Web { ISWC 2020 . pp. 430 { 446 . Springer International Publishing ( 2020 )

7. Wang , Q. , Mao , Z. , Wang , B. , Guo , L. : Knowledge Graph Embedding: A Survey of Approaches and Applications . IEEE Transactions on Knowledge and Data Engineering 29 ( 12 ), 2724 { 2743 ( 2017 ). https://doi.org/10.1109/TKDE. 2017 .2754499