Dear Wikidata developers and contributors,
I see that it is currently not possible to run federated queries from wikidata to other sparql endpoints. I understand why you are not allowing this in the general case. However, it would be nice to allow this in the special case.
Would a patch allowing limited known remote sparql endpoints to org.wikidata.query.rdf.blazegraph.WikibaseContextListener be possible?
Regards, Jerven Bolleman
Hi!
Dear Wikidata developers and contributors,
I see that it is currently not possible to run federated queries from wikidata to other sparql endpoints. I understand why you are not allowing this in the general case. However, it would be nice to allow this in the special case.
Would a patch allowing limited known remote sparql endpoints to org.wikidata.query.rdf.blazegraph.WikibaseContextListener be possible?
Technically, it is possible and shouldn't be very hard to do. We need to figure out which endpoints we want to allow.
On Thu, May 26, 2016 at 9:34 PM Stas Malyshev smalyshev@wikimedia.org wrote:
Hi!
Dear Wikidata developers and contributors,
I see that it is currently not possible to run federated queries from wikidata to other sparql endpoints. I understand why you are not allowing this in the general case. However, it would be nice to allow this in the special case.
Would a patch allowing limited known remote sparql endpoints to org.wikidata.query.rdf.blazegraph.WikibaseContextListener be possible?
Technically, it is possible and shouldn't be very hard to do. We need to figure out which endpoints we want to allow.
What do we need to take into account to figure this out? What are the constraints/concerns/etc? During the last weeks several people said they'd like to see this happen and I think we should explore this more if possible.
Cheers Lydia
Hi!
What do we need to take into account to figure this out? What are the constraints/concerns/etc?
Hmm... Looks like I maybe spoke too soon. While it is not a problem to make *WDQS* accept federated queries, it may be a problem to make federated queries actually *work*, because production machines do not seem to have direct access to the internet.
Which means if we want to make it work, we'd need to ask ops for one-off exceptions to firewall policies, which are usually frowned upon - they are hard to maintain and in general require extra effort and aren't really best practice.
Maybe some of the use cases might be better served by TPF server (https://github.com/blazegraph/BlazegraphBasedTPFServer) - this does not enable federated queries per se but enables to produce content that can be queried externally more easily.