Joachim,

I'd be interested in hearing how that goes. Fuseki(1|2) can just act as a SPARQL server, no associated application in the war. The shiro.ini can easily be used to block everything not associated with the dataset services. This is possibly the important use case at the moment. As you say, a full customizable end user interface would be a great enhancement for Fuseki and not trivial.

        Andy

On 21/05/15 17:30, Neubert, Joachim wrote:
Hi Andy,

Well, I had supposed the web app has access to (at least) the config.ttl data. 
So, it's not so easy ...

 From a data publishers point of view, the first option you mention would be 
clearly preferable. Repackaging a .war is before deployment not such a big 
thing. But when queries evolve (with changed data structures and/or changed 
requirements), repackaging and repdeploying the application after each change 
could cause some pain.

Re. the UI in general, I suppose it's a good idea to seperate end user and 
system management tasks. (I have not yet dived into spiro config, but probably 
such a separation would make it easier to hide sysadmin options from end users.)

Having a generic yet customizable end user interface would be a great 
enhancement for Fuseki - and, I suppose, it would require lots of work. 
Unfortunately, due to very limited skills in Java as well as in Javascript, 
I'll not be able to contribute actual code, but I'd happily contribute to in 
discussing conceptual ideas and testing.

Re-considering the options for now, I suppose I can use the approach described 
in http://zbw.eu/labs/blog/publishing-sparql-queries-live for providing 
customized forms for the ZBW endpoints. That should work for Fuseki1 as well as 
for Fuseki2, so it does not hinder migrating in any way.

Cheers, Joachim

-----Ursprüngliche Nachricht-----
Von: Andy Seaborne [mailto:a...@apache.org]
Gesendet: Donnerstag, 21. Mai 2015 15:17
An: users@jena.apache.org
Betreff: Re: AW: Fuseki2 - how to customize example queries?

Joachim,

That is going to take some code.

The interface a webapp - it happens to be served from the same server as the 
database though not much actually requires that (mainly a security issue).  The 
database engine and the webapp are relatively separate things, by design.  
There are a set of operations, under /$/ to support the webapp.

Two things occur to be:

1/ /$/"get customization"?dataset= to return a JSON struct of URLs to get 
various things.

2/ A small step might be to allow qonsole-config.js to also have a map from 
dataset to examples, not just a global block.

Stepping back a moment:

At the moment, the Fuseki2 UI has mixed purposes. Originally for server admin, the person 
or group publishing the data, there was a query interface (because the 
publisher/administrator might actually want to look at the data!).  That's why the 
dataset.html has "select endpoint"
and upload.

But it is also a query application, for more general use, a wider audience.  
And per endpoint / per dataset.

Seems to me that the two functions need to be provided differently.

A page like dataset.html for data mgt tasks for the publisher, and a page for 
data consumer.

The latter is open ended - starting from customizable query example through 
ever more data specific tasks.  It would be nice if there wasn't a big chasm 
between the two.

        Andy

On 21/05/15 10:47, Neubert, Joachim wrote:
Hi Andy,

Thanks for the hint. However, I can't see how this data structure
could
  > be used/extended to provide *dataset-specific* queries, as in e.g.
http://zbw.eu/beta/sparql/stw or http://zbw.eu/beta/sparql/gnd.

Could you imagine to supplement the generic queries with additional
custom queries loadedfrom a file referenced in config.ttl (opitonally, of 
course)?

Something like

<#someDataset>  rdf:type ja:RDFDataset ;
      skos:example <file:examples/someDataset.js>" .

(with someDataset.js  in the same syntax as qonsole-config.js).

Cheers, Joachim

-----Ursprüngliche Nachricht-----
Von: Andy Seaborne [mailto:a...@apache.org]
Gesendet: Donnerstag, 21. Mai 2015 00:05
An: users@jena.apache.org
Betreff: Re: Fuseki2 - how to customize example queries?

On 20/05/15 09:07, Neubert, Joachim wrote:
Congrats to the Fuseki2 interface - it's really fun to explore!

The generic example queries (Selection of triples, Selection of classes) can 
help giving users an idea about a dataset. They could be amended however with 
custom, dataset-specific queries (and prefixes relevant for that dataset).

Is there a way to do this?

They are in the file:

webapp/js/app/qonsole-config.js

If you are using the standalone jar: webapp is in the FUSEKI_HOME area

If you are using the war, you'll need to unpack the war and change it.

        Andy


Cheers, Joachim






Reply via email to