Identify specific issues for producing SPARQL according to PKG data model (utils.py)

[kbalog]

No description provided.

[mgskjaeveland]

Suggested pseudo algorithm for producing statements from incoming text.

1. Create a fresh rdf:Statement S
2. Add the incoming text as dc:description to S
3. Identify the subject, predicate and object of the statement.
4. For each X in (subject, predicate, object):
   • if X can be identified by a single IRI, then add 'S rdf:subject|rdf:predicate|rdf:object X'
   • else
     • create a skos:Concept C, and add X as a description to C
     • if related entities can be extracted from X and be identified by an IRI, then add these as 'skos:related|skos:broader|skos:narrower' to C
5. Add provenance data to to S (created by, ...)

An example for "I dislike all Tom Cruise movies" is found here: https://github.com/iai-group/pkg-vocabulary/blob/main/example/ex1.ttl