Open balhoff opened 6 years ago
What happens to these in the GPAD export pipeline (pre-loading into the main RDF store).
Is there a way to have Arachne provide some minimal sensible inferences in the face of inconsistencies (e.g. don't use rules that reference Nothing?).
The GPAD export will throw an InconsistentOntologyException
, and the command line should log a warning and skip the file.
Is there a way to have Arachne provide some minimal sensible inferences in the face of inconsistencies (e.g. don't use rules that reference Nothing?).
This can be done, although you might still get some weird results from domains and ranges.
There are quite a few inconsistencies in the RDF store: http://yasgui.org/short/HJjHiP1Az
The inferences produced for these models can be quite misleading, so I think we should prevent inferred graphs from being loaded into the RDF store for these models. This may require an update to blazegraph-runner as well some handling in the pipeline.
We should also have a means to notify the appropriate people to get these models fixed.