search

rubensworks / rdf-dereference.js

Dereference any URL for its RDF contents
MIT License
33 stars 4 forks source link

Fails to parse certain triples with JSON-LD #29

Closed brechtvdv closed 3 years ago

brechtvdv commented 3 years ago

This is probably an issue with the JSON-LD streaming parser, but I put the issue here for the reproducability.

I publish a LD document (https://lodi.ilabt.imec.be/coghent/dmg/objecten?generatedAtTime=2021-05-21T17:12:34.363Z) that contains information about subject https://stad.gent/id/mensgemaaktobject/dmg/530026423/2021-05-21T17:12:34.363Z and contains a link to its Production activity: image

Following triple should be parsed:https://stad.gent/id/mensgemaaktobject/dmg/530026423/2021-05-21T17:12:34.363Z http://www.cidoc-crm.org/cidoc-crm/P108i_was_produced_by _:b0 In the JSON-LD playground, this works: image

When I run rdf-dereference to use the RDF parser, I don't retrieve this triple:

rdf-dereference https://lodi.ilabt.imec.be/coghent/dmg/objecten?generatedAtTime=2021-05-21T17:12:34.363Z | grep "https://stad.gent/id/mensgemaaktobject/dmg/530026423/2021-05-21T17:12:34.363Z" | grep "produced"
rubensworks commented 3 years ago

Does it only occur when parsing this full document, or can it also be reproduced in a smaller subset of the document? (The latter would make debugging a lot easier)

brechtvdv commented 3 years ago

I made a subset of the document here: https://brechtvdv.github.io/demo-data/example-production-oslo.jsonld

The triple that I see missing is: 

<https://stad.gent/id/mensgemaaktobject/dmg/530026423/2021-05-21T17:12:34.363Z> <http://www.cidoc-crm.org/cidoc-crm/P108i_was_produced_by> _:b0