Ontotext Refine 1.0 Turns Strings into Things

Ontotext Refine takes over from GraphDB OntoRefine and becomes a standalone application with installers for all major OSs, a Docker image and configurable remote access to GraphDB

New York, Sofia, Basel Thursday, July 28, 2022

We are pleased to announce that as of GraphDB 10, the OntoRefine tool has been removed from Ontotext’s leading RDF database and is now an independent product – Ontotext Refine. Since 2017 and GraphDB 8.0, the open source tool OpenRefine has been integrated in GraphDB Workbench coupled with a powerful RDF mapping capabilities.

Ontotext Refine is an upgraded version of OpenRefine, specially tailored for the transformation of any machine readable data into RDF. OpenRefine’s large set of specialized data wrangling functionalities and visual interface helps users bring their input data to the quality required for RDF transformation. This can be anything from splitting several fields to manually homogenizing individual values using regular expressions and transposing columns into rows to even cross-referencing with a lookup table and reconciling with an external reconciliation endpoint, in order to map strings from the input data to things from a knowledge graph. OpenRefine excels at all of the above and adds functionalities such as a full history of all transformations (undo/redo), sweet data exploration functionalities (facets and filters) and the ability to collaborate on a single data cleaning project.

Once the tabular data is prepared, Ontotext Refine’s RDF visual mapper allows users to iteratively create a mapping specification and convert their tabular data to RDF one relation at a time. During this process, it allows the user to monitor the output (in turtle format) for possible errors and inconsistencies. The visual mapper generates a SPARQL mapping definition, which encodes the mapping as a construct or insert query. At the same time, Ontotext Refine exposes the data transformation project as a virtual SPARQL endpoint upon which the query could be executed. Finally if needed, this process can also be automatized using the ontotext-refine-cli tool as part of a more complex data processing pipeline.

This allows users not only to produce RDF with the SPARQL construct query but, more crucially, to use SPARQL federation to simultaneously query the Ontotext Refine virtual endpoint from any other endpoint, running over an existing knowledge graph. This federated mode enables users to exploit the full expressivity of SPARQL, all the while using the information in an existing knowledge graph to smoothly integrate it with the new data.

Ontotext Refine is free, available for download for all major operating systems as well as a platform independent distribution and a docker image.

Give Ontotext Refine a try!

New call-to-action


For more information, contact Doug Kimball, Chief Marketing Officer at Ontotext