How to map additional metadata fields in JSONLD @context.

[surchs]

I am trying to write a (JSONLD) query for information in the dataStandard field of AssetsSummary but noticed that dataStandard is not in the released @context. As far as I understand, this context is generated by generate_context which seems to skip fields that don't have a nskey attribute set. The dataStandard field does not have a nskey.

I think dataStandard could fit with the schema.org encodingFormat property (https://schema.org/encodingFormat). My questions are:

• is there a reason dataStandard is currently not in the context
• how could I contribute a suggested new mapping (e.g. to the sdo:encodingFormat), i.e. where would this information go?
• would you like me to do a PR for that?

[satra]

@surchs - a PR would be much welcome. for this perhaps we keep dataStandard, we do use encodingFormat for the actual files being used, while standards were meant to refer to a more meta organizational standard and could eventually include multiple.

the missing nskey was an oversight, so a PR would be good perhaps use nskey=dandi for now, unless you can think of a better property.

[surchs]

Ah, great. Thanks @satra, that makes sense. And I didn't realize that the attribute names are equivalent to their mapped vocabulary terms (plus the namespace). I'll make a PR for the dandi:dataStandard mapping.

I guess a follow-up question would be: since the context mapping seems to take the form of current_field_name + name_space_prefix, how would I be able to map e.g. the AssetSummary.dataStandard field to a term with a different name, e.g. example:dataType? Would we then change the dataStandard field to match the new mapping name?

[satra]

there are two ways of doing this. one is in the context file generation itself or building out a separate mapping component to the dandi namespace using sameAs or some such avenue. we have not decided on this yet, so any thoughts are welcome.

[surchs]

Thanks for the clarification, I'll give it a little think myself!