This repository contains code for the Elas4RDF Web Application. The code of the Question Answering Component is available at https://github.com/cnikas/elas4rdf_qa.
Elas4RDF is a scalable, efficient and effective method for keyword search over RDF datasets. It follows a triple-centered approach (in contrast to the entity-based approach) for providing more precise and explainable results. It relies on a special configuration of ElasticSearch for the needs of RDF. The approach is totally schema agnostic, and thus widely applicable. For capturing a wide range of information needs, Elas4RDF offers various perspectives for the search results. The current deployment is over DBpedia version 2015-10. The evaluation over DBpedia-Entity v2 test collection has shown that Elas4RDF performs similarly to systems built from scratch and use entity-oriented and dataset-specific index structures.
mvn clean install -DargLine="-Xmx4g" //use at least 4g ram
java -jar -Xmx4g target/elas4rdf-demo-0.0.1-SNAPSHOT.jar
OR
mvnw spring-boot:run //windows
./mvnw spring-boot:run //linux
Runs on: localhost:8081/elas4rdf
- qaurl: the url for the QA component's REST endpoint
- elas4rdfurl: the url of the instance of elas4rdf search service used
- datasetId: the dataset id on elas4rdf (e.g. dbpedia)
- extField: the field used for the extended information (e.g. rdfs_comment)