nfdi4plants / ARCtrl

Library for management of Annotated Research Contexts (ARCs) using an in-memory representation and runtime-agnostic contract systems.

MIT License

11 stars 8 forks source link

[BUG] RO crate WIP #326

Open Freymaurer opened 6 months ago

Freymaurer commented 6 months ago

Sources:

Must still be done:

Encoder

Assay

[x] Missing creator
[x] Missing identifier
[x] Missing measurementMethod
[x] Has not specified instrument

Covered by https://github.com/nfdi4plants/isa-ro-crate-profile/issues/9?

[ ] Has not specified url

Covered by https://github.com/nfdi4plants/isa-ro-crate-profile/issues/5?

[ ] Missing headline
[ ] Missing description
[ ] Missing dateCreated
[ ] Missing dateModified
[ ] Missing citation

LabProtocol

https://github.com/nfdi4plants/isa-ro-crate-profile/blob/main/profile/isa_ro_crate.md#labprotocol

[x] Missing reagent
[x] Missing computationalTool
[x] url has strange value "http://nfdi4plants.org/protocols/MyProtocol". Just a placeholder

Covered by https://github.com/nfdi4plants/isa-ro-crate-profile/issues/13?

[ ] Missing sameAs

Covered by https://github.com/nfdi4plants/isa-ro-crate-profile/issues/10?

[ ] Missing headline
[ ] Has not specified name

Sample

[x] processSequence.outputs has additionalProperty mapped to "characteristics" and "factorValues", instead of or.

Person

[ ] requires familiyName, but some cultures do not have one. PLUS if we must infer a person object from a isa-json authorList string we can only set name. I recommend setting familiyName as COULD. Therefore, we should also set givenName as SHOULD or move the inferred name from name to givenName.

ScholarlyArticle

[ ] sameAs is used for DOI in the example, but the given example doi is just a doi "11.1111/abcdef123456789", which is not of type URL.
[ ] url is used for PubMedID in the example, but the given example PMID is just a number "12345678", which is not of type URL.
[ ] i recommend setting both DOI and PMID as identifier of type PropertyValue with value as the identifier and valueReference as "DOI"/"PMID"/..etc

More

[x] Do we really need to switch between PropertyValue and DefinedTerm for ontology annotations? Deciding for either one seems like a better idea.

Decoder

Everything

HLWeil commented 6 months ago

Regarding the PropertyValue vs DefinedTerm: These are (in principle) not meant to be used interchangeably. Defined Term is used to depict a single ontology term. Property Value on the other hand is used to depict a term triplet (like parameterValue in isa). It can have an ontology header, ontology value and/or ontology unit.