How to use actor-query-process-annotate-source-binding #1396

balhoff · 2024-07-12T23:02:43Z

balhoff
Jul 12, 2024

I came across this actor, and it would be really useful to me to report the data sources for query solutions. I tried adding it to https://github.com/comunica/comunica/blob/master/engines/config-query-sparql/config/query-process/actors.json and found that it did start outputting _source bindings, but the value is always [], e.g.:

{"cell":"http://purl.obolibrary.org/obo/CL_0000182","cell_label":"\"hepatocyte\"","_source":"\"[]\""}

Is there further configuration needed to populate the _source binding?

Answered by RubenEschauzier

Jul 15, 2024

The actor-query-process-annotate-source-binding actor is only the last step to get the provenance of solutions from Comunica. You need to add a few more actors to compute the provenance information.

This PR aims to add these configs to Comunica link-traversal.

In summary, you need to add the following query-source-identify-hypermedia actor to your config:

    {
      "@id": "urn:comunica:default:query-source-identify-hypermedia/actors#annotate-source",
      "@type": "ActorQuerySourceIdentifyHypermediaAnnotateSource",
      "mediatorMergeBindingsContext": { "@id": "urn:comunica:default:merge-bindings-context/mediators#main" },
      "mediatorQuerySourceIdentifyHypermedia": { "@id": "urn:c…

View full answer

rubensworks · 2024-07-13T06:58:03Z

rubensworks
Jul 13, 2024
Maintainer

@RubenEschauzier Could you answer this one?

0 replies

RubenEschauzier · 2024-07-15T08:51:36Z

RubenEschauzier
Jul 15, 2024

The actor-query-process-annotate-source-binding actor is only the last step to get the provenance of solutions from Comunica. You need to add a few more actors to compute the provenance information.

This PR aims to add these configs to Comunica link-traversal.

In summary, you need to add the following query-source-identify-hypermedia actor to your config:

    {
      "@id": "urn:comunica:default:query-source-identify-hypermedia/actors#annotate-source",
      "@type": "ActorQuerySourceIdentifyHypermediaAnnotateSource",
      "mediatorMergeBindingsContext": { "@id": "urn:comunica:default:merge-bindings-context/mediators#main" },
      "mediatorQuerySourceIdentifyHypermedia": { "@id": "urn:comunica:default:query-source-identify-hypermedia/mediators#main" },
      "beforeActors": [
        { "@id": "urn:comunica:default:query-source-identify-hypermedia/actors#none" },
        { "@id": "urn:comunica:default:rdf-resolve-quad-pattern/actors#sparql" },
        { "@id": "urn:comunica:default:query-source-identify-hypermedia/actors#qpf" }
      ]
    }

This will add the source of the bindings to their context.

Then to ensure that during the join of bindings, the sources required for the result are the union of the two input binding sources, you need to add the following merge-bindings-context actor:

{
      "@id": "urn:comunica:default:merge-bindings-context/actors#source-binding-union",
      "@type": "ActorMergeBindingsContextUnion",
      "contextKey": "@comunica/bus-merge-bindings-context:sourceBinding"
}

Then the actor-query-process-annotate-source-binding actor will take the source information from the binding context and put it as an actual binding in the query result.

6 replies

RubenEschauzier Jul 17, 2024

No problem, let me know if any issues show up in the clean slate.
I would expect it to work on both TPF and SPARQL endpoints, but I haven't tested it specifically on that combination, so do let me know if any errors or unwanted behaviors appear.

balhoff Jul 23, 2024
Author

@RubenEschauzier here are the changes I made to try this out: master...balhoff:comunica:test-annotate-source

I added some logging to see if the new actors were running. When I run a query against two QPF endpoints, I get these lines printed (you can see where I put console.log in the diff):

ActorQueryProcessAnnotateSourceBinding
ActorMergeBindingsContextUnion
ActorMergeBindingsContextUnion
ActorQuerySourceIdentifyHypermediaAnnotateSource - TEST
ActorMergeBindingsContextUnion
ActorMergeBindingsContextUnion
ActorQuerySourceIdentifyHypermediaAnnotateSource - TEST
ActorMergeBindingsContextUnion
ActorMergeBindingsContextUnion
ActorMergeBindingsContextUnion

The run method of ActorQuerySourceIdentifyHypermediaAnnotateSource never seems to be called, and the _source binding is not populated. Am I missing something? Thank you!

RubenEschauzier Aug 7, 2024

This might be a bug that occurs when annotating QPF endpoints. Are the QPF endpoints available publicly so I can reproduce the problem locally?
(Note I'll be on leave again from 13-25 August, so responses will be somewhat delayed in that period)

balhoff Aug 7, 2024
Author

@RubenEschauzier yes they are public, here is an example query:

node bin/query.js 'https://frink.apps.renci.org/ldf/wikidata' 'https://frink.apps.renci.org/ldf/ubergraph' 'PREFIX wd: <http://www.wikidata.org/entity/>
PREFIX wdt: <http://www.wikidata.org/prop/direct/>
PREFIX wdtn: <http://www.wikidata.org/prop/direct-normalized/>
PREFIX rdfs: <http://www.w3.org/2000/01/rdf-schema#>
PREFIX cell: <http://purl.obolibrary.org/obo/CL_0000000>
PREFIX abdomen: <http://purl.obolibrary.org/obo/UBERON_0000916>
PREFIX part_of: <http://purl.obolibrary.org/obo/BFO_0000050>
SELECT DISTINCT ?cell ?cell_label ?article ?title
WHERE {
  ?cell rdfs:subClassOf cell: .
  ?cell part_of: abdomen: .
  ?wikicell wdtn:P7963 ?cell .
  ?article wdt:P921 ?wikicell .
  ?article rdfs:label ?title .
  ?cell rdfs:label ?cell_label .
}
LIMIT 20'

No hurry! And thank you for all your help.

RubenEschauzier Aug 26, 2024

Update: The problem is a bug with ActorQuerySourceIdentifyHypermediaAnnotateSource. As usual, fixing this revealed another bug. Both should be fixed in #1420, which (hopefully) only requires slight polishing before being merge-ready.

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Uh oh!

How to use actor-query-process-annotate-source-binding #1396

Uh oh!

{{title}}

Uh oh!

Replies: 2 comments 6 replies

Uh oh!

{{title}}

Uh oh!

Uh oh!

{{title}}

Uh oh!

Uh oh!

{{title}}

Uh oh!

Uh oh!

{{title}}

Uh oh!

Uh oh!

{{title}}

Uh oh!

Uh oh!

{{title}}

Uh oh!

Uh oh!

{{title}}

Uh oh!

Select a reply

Uh oh!

Uh oh!

How to use actor-query-process-annotate-source-binding #1396

Uh oh!

balhoff Jul 12, 2024

Replies: 2 comments · 6 replies

Uh oh!

rubensworks Jul 13, 2024 Maintainer

Uh oh!

RubenEschauzier Jul 15, 2024

Uh oh!

RubenEschauzier Jul 17, 2024

Uh oh!

balhoff Jul 23, 2024 Author

Uh oh!

RubenEschauzier Aug 7, 2024

Uh oh!

balhoff Aug 7, 2024 Author

Uh oh!

RubenEschauzier Aug 26, 2024

balhoff
Jul 12, 2024

Replies: 2 comments 6 replies

rubensworks
Jul 13, 2024
Maintainer

RubenEschauzier
Jul 15, 2024

balhoff Jul 23, 2024
Author

balhoff Aug 7, 2024
Author