naas.ai Docs

Triple-Store

1 min read

Triple Store Service

TripleStoreService is the RDF and SPARQL core for ontologies and semantic data.

Related pages: Engine, Bus-Service, Built-in-Module-Templatable-SPARQL.

Core capabilities

  • Insert/remove RDF graphs.
  • Query with SPARQL (SELECT, CONSTRUCT, etc.).
  • Subject-level graph retrieval.
  • Named graph operations (adapter-dependent).
  • Schema loading with update detection (load_schema, load_schemas).
  • Event publication to bus on insert/delete (when services are wired).

Adapter options

  • fs: filesystem-backed local triple partitioning by subject hash.
  • object_storage: object-storage-backed partitioning.
  • oxigraph: HTTP adapter with named graph support.
  • apache_jena_tdb2: Fuseki/TDB2 HTTP adapter with named graph support.
  • aws_neptune: direct AWS Neptune adapter.
  • aws_neptune_sshtunnel: Neptune via SSH tunnel.
  • custom: pluggable custom adapter.

Important compatibility notes

  • Filesystem and object-storage adapters do not support named graphs (create_graph, drop_graph, etc.).
  • Oxigraph/Jena/Neptune adapters support named graph management.
  • Triple insert/remove emits bus events using hashed routing keys.

Minimal usage

from rdflib import Graph
 
triples = Graph()
# ... add triples
 
engine.services.triple_store.insert(triples)
result = engine.services.triple_store.query("SELECT ?s ?p ?o WHERE { ?s ?p ?o } LIMIT 10")

Schema management behavior

load_schema(path) tracks metadata in an internal ontology:

  • file path
  • file hash
  • file modified time
  • base64 content

When a schema changes, it computes delta triples and applies insert/remove updates.

Graph View