quantities / PID for units

[jmilesBIPM]

When generating the file quantities.ttl, please could you include the PID for the unit each time?
(You can use the API /quantities/{parameter} to obtain the PID.)

Thanks!

[fmeynadier]

Hm, I thought it was straightforward but it is not....

Currently the si:hasUnit predicate already points to a PID if the unit is simple, or a to a CompoundUnit (or is absent in case the unit is "null" in the YAML), e.g.

quantities:ABDO a si:QuantityKind ;
    skos:altLabel "ABDO"^^xsd:string ;
    skos:prefLabel "absorbed dose"@en,
        "dose absorbée"@fr ;
    si:hasUnit units:gray .

quantities:ABDR a si:QuantityKind ;
    skos:altLabel "ABDR"^^xsd:string ;
    skos:prefLabel "absorbed dose rate"@en,
        "débit de dose absorbée"@fr ;
    si:hasUnit [ a si:UnitProduct ;
            si:hasLeftUnitTerm units:gray ;
            si:hasRightUnitTerm [ a si:UnitPower ;
                    si:hasNumericExponent "-1"^^xsd:short ;
                    si:hasUnitBase units:second ] ] .

quantities:ANIS a si:QuantityKind ;
    skos:altLabel "ANIS"^^xsd:string ;
    skos:prefLabel "anisotropy"@en,
        "anisotropie"@fr .

In the case of the compound units one could imagine to point to the PID (instead of the CompoundUnit ? Or as a separate relationship - but then it would pose a problem of information duplication and possible mismatch, even for simple units ?). And for "null" units should a "null unit PID" be created ?

I think the approach so far was that it would be the user's job to derive a PID from the CompoundUnit if needed, the algorithm to do that is non-ambiguous I think.

[jmilesBIPM]

For "null" units we're good: there's no si:hasUnit.

For simple units, we're pretty good: we give the PID through reference to the units file (but we could alternatively provide the PID in full).

For compound units we provide a unit expression but we're missing its PID... To improve consistency and avoid duplication, my suggestion would be to provide the full PID in both cases.