On possibility of updating to RDF4J

[schivmeister]

I was wondering why, instead of updating to RDF4J, Pinto was only updated to Sesame 4.x? Was this a conscious decision based on certain things or just a preference?

http://docs.rdf4j.org/migration/

[mhgrove]

just a preference, mostly because that's the version we've standardized on

[schivmeister]

I don't think it's that much work from 4.x, but I won't know until I dig. It looks like namespace changes mostly, but there could be other regressions waiting. But you have a nice test suite already that should cover that.

By any chance, if I chose to pursue an update myself, I won't be able to do it since your Sesame dependencies are coming from Stardog repositories, correct?

[mhgrove]

no, this is not used in stardog, so it could be upgraded without affecting anything else

[schivmeister]

@mhgrove So, I finally got to this, but I am not sure if https://github.com/mhgrove/cp-openrdf-utils is up-to-date or there are some internal changes that you didn't make public. The latest version there is 4.0.1 but here it's 4.0.2.

After updating the namespaces to RDF4J, the build and all but two tests pass in there. And here, build fails because Models2.toList() is having an incompatible signature (List<Value>, Model) vs. (List<Resource>). UUIDCodec.writeValue() also has an incompatible return type ResourceBuilder vs. Value.

There appear to be no other errors (there was one wrt URI, but I changed that to IRI and it became happy). So I think it's not too much work to update to RDF4J as this looks close. I can send a pull request for both projects if you can help clarify those remaining issues.

[mhgrove]

not sure what the version mismatch is for the openrdf utils, but probably not a big deal. the migration sounds good otherwise

[schivmeister]

mhgrove/cp-openrdf-utils#23