The -base release products (metpo-base.owl / .obo / .json) ship with zero METPO terms (only ontology metadata + annotation-property declarations). The full artifacts are fine.
Verified:
metpo-base.owl on main and in the 2026-05-19 release: 0 owl:Class for https://w3id.org/metpo/<id>.metpo.owl / metpo-full.owl: 1454 METPO classes.
So this is pre-existing, across multiple past releases; it was not introduced by #462 (that PR's URIBASE change moved one non-matching base-IRI to another non-matching one).
Root cause: the base pipeline in src/ontology/Makefile does
remove --base-iri $(URIBASE)/METPO --axioms external ...
but METPO term IRIs are https://w3id.org/metpo/<id> (lowercase, slash-separated). $(URIBASE)/METPO never prefix-matches a real term IRI, so remove --axioms external strips every METPO axiom from the base. The QC report rules (lines 245/248) already use the correct $(URIBASE)/metpo.
Fix: change line 542 to --base-iri $(URIBASE)/metpo and rebuild so the base carries the 1454 METPO terms minus imported externals.
Surfaced by Copilot's review on #462. Related: #436, #433.
The
-baserelease products (metpo-base.owl/.obo/.json) ship with zero METPO terms (only ontology metadata + annotation-property declarations). The full artifacts are fine.Verified:
metpo-base.owlonmainand in the2026-05-19release: 0owl:Classforhttps://w3id.org/metpo/<id>.metpo.owl/metpo-full.owl: 1454 METPO classes.So this is pre-existing, across multiple past releases; it was not introduced by #462 (that PR's
URIBASEchange moved one non-matching base-IRI to another non-matching one).Root cause: the base pipeline in
src/ontology/Makefiledoesbut METPO term IRIs are
https://w3id.org/metpo/<id>(lowercase, slash-separated).$(URIBASE)/METPOnever prefix-matches a real term IRI, soremove --axioms externalstrips every METPO axiom from the base. The QC report rules (lines 245/248) already use the correct$(URIBASE)/metpo.Fix: change line 542 to
--base-iri $(URIBASE)/metpoand rebuild so the base carries the 1454 METPO terms minus imported externals.Surfaced by Copilot's review on #462. Related: #436, #433.