Dear Katerina,
thanks for reporting on the version you are using. I strongly suggest to move to the new VB 15.1 and SV 6.1 as, besides several improvements exactly on metadata (thanks to a new version of the MDR vocabulary exploiting almost all features of DCAT3 and bringing it together with other relevant vocabularies such as VoID and LIME), it offers support for multiple versions (with metadata for each of them), editable forks in VB etc.. and they also removed that endpoint field you would like not to show :-) (but, in the MDR, you can then represent the official SPARQL endpoint and/or endpoints for the various versions of a dataset, if available)
Kind Regards,
Armando
From: vocben...@googlegroups.com <vocben...@googlegroups.com> On Behalf Of Katerina Gkirtzou
Sent: Thursday, April 30, 2026 4:05 PM
To: vocbench-user <vocben...@googlegroups.com>
Subject: [vocbench-user] Showvoc shows docker internal address as sparql endpoint
Dear community,
We have set up an installation of vocbench and showvoc, along with their respective external graphdb for our needs via Docker Compose. Furthermore, we opted to hide the different ports behind nginx (we successfully managed to hide 3 out of the 4 ports) along with an SSL certificate for security and accessibility reasons.
In showvoc, we have set the Remote access configuration-> Server UR with: http://showvoc-graphdb:7200/, which is the Docker internal address, following the instructions you provided for setting up showvoc via Docker (https://bitbucket.org/art-uniroma2/showvoc-docker/src/master/).
We've just realised that this information is shown at the public metadata page of our vocabularies under the SPARQL Endpoint label (see screenshot). Is there a way to hide this information from showing, as this is internal information for Showvoc? Furthermore, the graph_db is password-protected, and in theory, it is not accessible.
Let me clarify that we are using the following versions:
SHOWVOC_VERSION="5.0.0"
ST_VERSION="14.0"
VB_VERSION="14.0.0"
GDB_VERSION="10.6.2"
Thanks in advance for your help.
Kind regards,
Katerina
--
You received this message because you are subscribed to the Google Groups "vocbench-user" group.
To unsubscribe from this group and stop receiving emails from it, send an email to vocbench-use...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/vocbench-user/556b257f-630c-4531-bdb3-5b76d3ffbb42n%40googlegroups.com.
Dear Armando,
thank you for your
rapid response. I will checkout the newest version of VB and SV,
to see all these latest improvements and if indeed helps us hide
the internal SPARQL endpoint of docker address. I do have some
questions though. You said that "in the MDR, you can then represent
the official SPARQL endpoint and/or endpoints for the various
versions of a dataset, if available". Is it possible to use
the showvoc (and it's internal semantic turkey) as the
official SPARQL endpoint for our published vocabularies via
showvoc? If yes, do you have any documentation which documents
such mechanism?
Secondly, as we use SV and VB in a production deployment
setting for publishing our vocabularies, before upgrading to
the latest versions I need to verify that none of the data in
our system is lost. So my question is how should I backup
the information stored in the semantic turkey for both VB and SV
(projects, user, configurations, etc) ? Is it sufficient to copy
the folders where respective volumes live, as we use
docker-compose to deploy them? Note that the majority of the
actual data are stored in the related graphdb, so the system
points to external db for the most projects. Finally, are the
latest versions backward compatible with the versions we use
(SV, VB, ST, graphdb)?
Thank in advance for all.
Kind regards,
Katerina
Dear Katerina,
the SPARQL interface of VocBench and ShowVoc is, on purpose, restricted, passing through their own API. This is because leaving a SPARQL endpoint open is a too dangerous operation, definitely for write operation (ok, that could be limited) but also for accessing the content, as several SPARQL queries run by “anyone from anywhere” could definitely put a strain on the system.
Besides this, there is one reason for which we didn’t, in any case, foresee a free SPARQL access based on standard GraphStore/SPARQL API: assuming that, as recommended, in production you would use a separate triple store, then it would be much better to configure the triplestore as an endpoint and there would be no purpose nor gain in putting VB/SV in the middle.
Just for completeness: the SPARQL API for VB/SV are quite close (same parameters) to the standard API, just they include the additional aspects of authentication/authorization provided by VB/SV.
About the backup, that is very easy:
Hope it helps!
Armando
Da: vocben...@googlegroups.com <vocben...@googlegroups.com> Per conto di Katerina Gkirtzou
Inviato: mercoledì 6 maggio 2026 10:11
A: stel...@uniroma2.it; 'vocbench-user' <vocben...@googlegroups.com>
Oggetto: Re: [vocbench-user] Showvoc shows docker internal address as sparql endpoint
To view this discussion visit https://groups.google.com/d/msgid/vocbench-user/19afe383-9f37-48b1-bb69-92f4ef24f3e0%40gmail.com.
Dear Armando,
thanks for your prompt and detail response. It is indeed very helpful and provides clarity. I think that I have enough information regarding bakcup and upgrade, but if I need anything I will let you know.
Thanks again.
Kind regards,
Katerina