Open tpluscode opened 3 years ago
@tpluscode I have been working on this... I have made a lot of progress but I am not done. There are several intersections between types used in sparql-engine
, n3
, and sparqljs
, and not only do the typings change as you push up to higher versions of each library, but the API's as well, so my progress has been slow. Since sparql-engine
transitions between hard typings moving in and out of the pipeline and some expanded types returned by sparqljs
from SPARQL* and property path queries, the algebra needed to be streamlined too. But one of the first things I did was create a namespace for sparql-engine
apart from the existing `sparqljs' namespace and I have all types descending from @types/rdf-js.
Hi,
After a long hiatus, I've finally had time to dedicate myself to Open Source again.
I agree with @tpluscode : it is a shame that sparql-engine
doesn't align with the RDF/JS standard. For context, we developed this framework when RDF/JS was not in a mature state. However, since then, we have migrated some parts of the framework internals classes to work with RDF/JS natively, but we didn't finish the migration.
As of today, I'm taking care of this issue and I've started working on a complete migration from our native RDF model to RDF/JS spec. I will update this thread as progress are made, but if you have any suggestion, feel free to comment! š
EDIT: All work will be pushed to the develop/1.0.0_rdfjs branch
@Callidon Are you still working on that or is the issue stalled? Don't see much activity. I agree that RDFJS support would be very helpful, in that form it's not that useful as we can't integrate the library into the existing RDFJS stack.
Hi,
I agree it is a good topic, but my current work activity doesn't allow me to focus on such side project anymore. Soi sadly, unless someone is willing to work on it, the issue is going to stale. The branch develop/1.0.0_rdjs contains a good start on what I've done until now, but realistically, I cannot allow more time to it. Sorry.
@ktk @Callidon I had worked out much of the changeover in a fork I was working on but then got slammed with other projects. I want to finish my efforts on this but cannot return to it yet. Unfortunately, this is a difficult task to accomplish incrementally, but I believe this implementation of SPARQL is one of the best I have seen in JavaScript, so I believe it is worth the effort. I cant make any promises either, but I would be willing to work with someone else if they were to invest some time too.
@JuniperChicago is that fork public? We can have a look at it @zazuko, maybe we can work it out together.
@JuniperChicago are you still working on it, or is the fork public as @ktk asked?
I'm also very interested in this, since I need SHACL SPARQL support in rdf-validate-shacl (e.g. see. https://github.com/zazuko/rdf-validate-shacl/issues/104).
Is your feature request related to a problem? Please describe.
I do not find any mention of RDF/JS spec in the docs. It would be really useful if the query engine was usable with any RDF/JS DatasetCore. It's a shame that the library uses its own abstraction of the RDF model.
Describe the solution you'd like
Ideally, the usage would be something like
Describe alternatives you've considered
An adapter would also be an option like below, but I think that given the incompatibility of the graph and triple models this would not make a great solution
Additional context
RDF/JS is a de-facto standard, driven by community, and widely supported. N3 is compatible, graph.js is compatible, comunica is compatible. I think it beneficiary to also use RDF/JS as the underlying model for sparql-engine