reasoning and inferencing in SPARQL endpoint

[carueda]

The ORR uses (a no-cost version of) https://allegrograph.com/ to power its triple-store and SPARQL endpoint.

Although AG does support some common reasoning and inferencing mechanisms this feature has not been enabled by default in the ORR, mainly due to rather poor performance behavior that was noted when these capabilities were tested a long time ago.

The task here is to revisit this aspect, initially with AG (including an update to its latest version) as well as investigate alternative triple stores, ideally open-source.

@lewismc Lewis, any other aspects to consider here?

[lewismc]

@carueda I totally missed this. I think we should focus on demonstrating this at the next COR telecon. What have we got to loose?
From memory, can you provide details on the following

• how long did it take
• what were the commands you had to execute on the server?
• what was the output?

I feel that we should run reasoning and inference plugins and then analyze the results. This would be beneficial from a SWEET perspective regardless (could also be included in the SWEET paper).