Frequently Asked Questions

  1. The quality of reconciliation results is poor. How can I enhance it?

    If the reconciliation service is defined as a general SPARQL one (i.e. using regular expressions for comparison) or if the data behind the endpoint is heterogeneous, the reconciliation results might be poor. You might consider running a SPARQL construct query against the endpoint to extract the data (or better the relevant subset of it). The result of the construct query is RDF data that can be used to define a RDF dump-based reconciliation service (as in this example)

  2. How do I get the reconciled URI in the RDF exporter?

    using the GREL expression cell.recon.match.id

    Notice that you can also get the values of the first candidate via the expression: cell.recon.candidates[0].id

  3. Can I use the extension in an automatic way (via a scheduled task for example)?

    In principle yes. However, you need to prepare some script yourself. The Python client library for Google Refine can be very helpful. Basically, you need to extract the operations history into a JSON file (including the RDF skelteon design, apply it to new files/data and then export using format='rdf' or format='ttl' (see this script).

  4. Is reconciliation using Silk Server still supported?

    Currently no. The support that is described in LDOW2011 paper was removed from the extension as we found no clear usecase or need for it. Interested in it? get in touch please.