Replace Allotrope Foundation (AFO) dependencies with open OBO/QUDT alternatives

[StroemPhi]

Affected: chemical_entities_ap.yaml (AFR:0002409 for MolarMass), chemical_reaction_ap.yaml (AFE:0000153 for Reactor)
Severity: Strategic - OBO alignment concern
Status: Requires community coordination

Problem: The Allotrope Foundation ontologies (AFR, AFE, AFP) are not really that open and their OBO alignment is questionable (see https://doi.org/10.1515/pac-2021-2007 for more). ChemDCAT-AP currently depends on AFO terms for MolarMass (AFR:0002409) and Reactor (AFE:0000153). This creates a questionalbe dependency in an otherwise fully open stack.

Proposed approach: Replace class_uri mappings with terms from NFDIcore, OBI, or CHMO. For quantity-typed classes that currently lack OBO equivalents, propose defined classes to the NFDIcore or OBO community following this pattern:

???:TemperatureDatum ≡ IAO:0000109 (measurement datum)
    and (IAO:0000136 some PATO:0000146)
    and (qudt:hasQuantityKind value qudt:Temperature)

This connects OBO (PATO quality, IAO datum) with QUDT (quantity kind) in a single defined class, replacing the AFO dependency with open alternatives. Although using IAO:0000109 (measurement datum) would be less suitable than just data item, to also include data items about qualities that are the result of simulations or just settings.

Context: Chris Mungall has mapped PATO qualities to QUDT QuantityKind via SSSOM. The correspondence is established; what's missing is OWL classes for the measured/recorded information about those qualities. QUDT provides the vocabulary (QuantityKind is a SKOS concept scheme) but not the ontological formalism. PATO provides OWL classes but for qualities, not information entities. The defined class pattern above bridges this gap.

Blocked by: NFDIcore collaboration progress, OBO community discussion.

[markdoerr]

@StroemPhi : I fully support avoiding AFO terms/dependencies out of the reasons you mentioned.