Remove graph limitations on `/describe/` service used in FCT?

Hello, i need advices to spot entry point in source code for the implementation of the /describe point of access in VOS.

The /describe service is used by the FCT (facet) VAD package, and explicitly limiting the triplets in the result to only one graph when you click the second tab (“Description”) in the facet webapp about a resource.

We know it’s possible to enlarge the results to other datasets (graphs) by providing a list of graph names in http parameters (as in Faceted Browser not listing all resources on page description) but this solution is unapplicable for us because we got tons of dynamically generated graphs (purpose is to “group” triples about a resource in distinct contexts)

The default SPARQL describe operator behaviour is to allow to query ALL graphs by default, we can illustrate with these two links :

  1. results send to the faceting system (only one specific graph) by

https://data.idref.fr/describe/?url=http://www.idref.fr/169814289/id

  1. results send by the “standard” SPARQL describe query, default behaviour is searching in all graphs of the triplestore :

https://data.idref.fr/sparql?query=DESCRIBE%20<http://www.idref.fr/169814289/id>&format=text/plain

= more triples from other graphs.

I suspect source code in /describe to limit explicitly the query to one specific graph.

Any help, clues to search in source code, advices, … are welcome,

Thanks !

Thomas

I note that your endpoint is running a rather old v1.13.54 of this VAD (dating from 2014 or earlier) while the latest pre-compiled is 1.13.91. (You can get a full complement of current VADs for VOS 7.2 here.)

It appears that your concern was resolved in v1.13.55 of this VAD, so unless there’s been some regression (which we would look into), you should be set by installing v1.13.91 or later.

That said, searching on "1.13.91" leads to develop/7/binsrc/b3s/vad_version, and in turn to the sourcecode I think you’re looking for….

Hello, release 1.13.91 of fct solved our problem ! We really appreciated your help, thanks a lot !