apunkt/trustgraph
1
0
Fork 0
mirror of https://github.com/trustgraph-ai/trustgraph.git synced 2026-04-25 16:36:21 +02:00
Code Issues Projects Releases Packages Wiki Activity Actions 1

Remove schema:subjectOf edges from KG extraction (#695)

Browse source 
The subjectOf triples were redundant with the subgraph provenance model
introduced in e8407b34. Entity-to-source lineage can be traced via
tg:contains -> subgraph -> prov:wasDerivedFrom -> chunk, making the
direct subjectOf edges unnecessary metadata polluting the knowledge graph.

Removed from all three extractors (agent, definitions, relationships),
cleaned up the SUBJECT_OF constant and vocabulary label, and updated
tests accordingly.
This commit is contained in:
cybermaggedon 2026-03-13 12:11:21 +00:00 • committed by GitHub
parent 64e3f6bd0d
commit e6623fc915
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
10 changed files with 9 additions and 88 deletions
Download patch file Download diff file Expand all files Collapse all files

16
tests/unit/test_knowledge_graph/test_agent_extraction.py
View file

@ -13,7 +13,7 @@ from unittest.mock import AsyncMock, MagicMock, patch
from trustgraph.extract.kg.agent.extract import Processor as AgentKgExtractor
from trustgraph.schema import Chunk, Triple, Triples, Metadata, Term, Error, IRI, LITERAL
from trustgraph.schema import EntityContext, EntityContexts
from trustgraph.rdf import TRUSTGRAPH_ENTITIES, DEFINITION, RDF_LABEL, SUBJECT_OF
from trustgraph.rdf import TRUSTGRAPH_ENTITIES, DEFINITION, RDF_LABEL
from trustgraph.template.prompt_manager import PromptManager
@ -183,12 +183,6 @@ This is not JSON at all
assert def_triple.s.iri == f"{TRUSTGRAPH_ENTITIES}Machine%20Learning"
assert def_triple.o.value == "A subset of AI that enables learning from data."
# Check subject-of triple
subject_of_triple = next((t for t in triples if t.p.iri == SUBJECT_OF), None)
assert subject_of_triple is not None
assert subject_of_triple.s.iri == f"{TRUSTGRAPH_ENTITIES}Machine%20Learning"
assert subject_of_triple.o.iri == "doc123"
# Check entity context
assert len(entity_contexts) == 1
assert entity_contexts[0].entity.iri == f"{TRUSTGRAPH_ENTITIES}Machine%20Learning"
@ -228,10 +222,6 @@ This is not JSON at all
assert rel_triple.o.iri == object_uri
assert rel_triple.o.type == IRI
# Check subject-of relationships
subject_of_triples = [t for t in triples if t.p.iri == SUBJECT_OF and t.o.iri == "doc123"]
assert len(subject_of_triples) >= 2 # At least subject and predicate should have subject-of relations
def test_process_extraction_data_literal_object(self, agent_extractor, sample_metadata):
"""Test processing of relationships with literal objects"""
data = [
@ -274,10 +264,6 @@ This is not JSON at all
triples, entity_contexts, _ = agent_extractor.process_extraction_data(data, metadata)
# Should not create subject-of relationships when no metadata ID
subject_of_triples = [t for t in triples if t.p.iri == SUBJECT_OF]
assert len(subject_of_triples) == 0
# Should still create entity contexts
assert len(entity_contexts) == 1