Support Multiple Datasets

[no-reply]

Allow multiple datasets in the cache service. The use case I have in mind for this is hosting multiple "authorities" in a single cache service.

Datasets are defined by the LD Fragments spec. We implement it in the Dataset model.

We should be able to introduce new datasets to act as additional subject cache endpoints. Ideally, it would be possible to put datasets in named graphs on a shared backend, or put them in separate backends.

[no-reply]

Started work on this.

[no-reply]

So, this is the point where I think I need to rework the config and/or the routes. I'm working on the most backward compatible ways to do this, but expecting this to require breaking changes. More info to come.

cc: @scande3 @elrayle

[no-reply]

Initial work on this is here: https://github.com/ActiveTriples/linked-data-fragments/tree/feature/multi-dataset

I'm more confident that routing shouldn't be a compatibility issue, at this point; or if it is, it will be very minor. Config is still likely a problem.

[no-reply]

https://github.com/ActiveTriples/linked-data-fragments/tree/feature/multi-dataset is updated with the latest work. A few PRs have been submitted representing some of that progress.

[no-reply]

https://github.com/ActiveTriples/linked-data-fragments/tree/feature/multi-dataset is a finished branch with this feature. Two things need to happen in order for this to reach PR:

• Reduce the VCR footprint.
  • Some of this should be manageable by using named cassettes; some of it may require refactoring.
• Get the Blazegraph & Marmotta backends passing the backend shared examples.
  • These didn't pass the relatively simple examples merged previously, and the interface has drifted a bit out from under them. I'll do the Marmotta work in the next week-ish; blazegraph may take slightly longer.

[hackartisan]

@no-reply Is Marmotta work referenced in this last comment still outstanding? I assume the blazegraph work is.

Is it obvious what this work is by running the tests on the branch?