Q: How to use the SPARQL endpoint to query EDG asset collections

[Tim Smith]

Hi,

As I attempt to design a linked data environment, I am considering how to use EDG.  A SPARQL endpoint is critical to such an environment.  However, using the TBC version of EDG (maybe the server version behaves differently), when I query the SPARQL endpoint using the External Graph URI as defined in the asset collection, no results are returned.  If I use the internal graph URI from EDG, results are returned.

Should the SPARQL endpoint recognize the external URI?  

Thanks,

Tim

[Irene Polikoff]

Hi Tim,

No.

External Graph URI setting is only for minimizing the effort for loading into EDG content that is available as a set of interconnected files with owl:imports statements.  And to (if needed) export back from EDG as a set of files with the base URIs specified in this setting. It is not used by the SPARQL endpoint. It is not used by any capability in EDG except for the RDF file import and export.

Regards,

Irene

--
You received this message because you are subscribed to the Google Groups "TopBraid Suite Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to topbraid-user...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/topbraid-users/CAF0WbnKxG9eVN9JaqHnGj_Lm36Cw%2B%3D9vF788OkUtvPqKCBYgRw%40mail.gmail.com.

[Tim Smith]

Hi Irene,

Ok, I understand.

What is TQ's vision for how EDG fits in a linked data ecosystem?  How does EDG share it's knowledge?  Off the top of my head, I would like to be able to:

1. Query the SPARQL endpoint to explore the graphs from external systems.  Do I need to export the graphs to set the external URIs and then load them into the TBL workspace?
2. Type the URL of a class or instance into my browser and have an SWP page (or some other method) popup showing navigable details of that class or instance.

Thanks,

Tim

To view this discussion on the web visit https://groups.google.com/d/msgid/topbraid-users/7387B524-1AE4-4A85-839F-49114442E2A2%40topquadrant.com.

[Irene Polikoff]

Hi Tim,

Please see below. 

On Apr 16, 2020, at 2:47 PM, Tim Smith <smith...@gmail.com> wrote:

Hi Irene,

Ok, I understand.

What is TQ's vision for how EDG fits in a linked data ecosystem?  How does EDG share it's knowledge?  Off the top of my head, I would like to be able to:

1. Query the SPARQL endpoint to explore the graphs from external systems.  Do I need to export the graphs to set the external URIs and then load them into the TBL workspace?

Graphs in EDG have graph URIs assigned to them by EDG. I would advise using these URIs in your queries. URI of resources in a graph are fully under your control. 

Typically, a graph to query is specified in the ?default-graph-uri parameter of the service call.

If, for some reason, you need to use a graph URI in a query that is different from graph's identity in EDG, then some intermediate layer would be needed to change the query or re-write the URI. 

Some examples of SPARQL endpoint access are here https://doc.topquadrant.com/6.3/developer-guide/#SPARQL_endpoint

2. Type the URL of a class or instance into my browser and have an SWP page (or some other method) popup showing navigable details of that class or instance.

Is this for a person or for an API? For API access EDG offers GraphQL query point (to return JSON of whatever structure you need) as well as SPARQL query point.

If it is for people, then you need some URI re-wrting rules at the Tomcat level to change for example:

http://mycompany.com/ns/examples/geography#Asia

To

http://<myedgserver>::<port>/tbl/geo.editor#http:%2F%2Fmycompany.com%2Fns%2Fexamples%2Fgeography%23Asia

I am not a Tomcat person, but, as I understand it, this is typically done through Apache HTTPD mod_rewrite (or something similar) using regular expression rules to make the rewrite work. It is not anything specific to TopBraid EDG, but rather something that is very commonly done with web technologies.

Thanks,

Tim

To view this discussion on the web visit https://groups.google.com/d/msgid/topbraid-users/CAF0Wbn%2BXS7f_ohVn%3DxDHnzHie5_5dGNLa%2Bf-BJSSQnOdsnSpBQ%40mail.gmail.com.

[Irene Polikoff]

And, of course, you need to own http://mycompany.com/ domain and have a web server registered at this address that will receive the HTTP requests for these URLs and will then re-write and re-route requests accordingly.