search

iodepo / odis-arch

Development of the Ocean Data and Information System (ODIS) architecture
https://book.oceaninfohub.org/
26 stars 16 forks source link

including Taxon type #391

Open sformel-usgs opened 5 months ago

sformel-usgs commented 5 months ago

I'm not sure if I'm in the right repo, feel free to point me elsewhere. I was surprised to find that ODIS doesn't include the Taxon type from schema.org (https://schema.org/Taxon). I gather that this is related to the thematic constraints of ODIS, but I couldn't find any other issues discussing the Taxon type. Are there plans to include it eventually?

Use case: I discovered this by searching for Cnidaria, and the results didn't include the NOAA Deep Sea Corals Research and Technology Program, a major resource for deep sea corals. In the OBIS resource metadata, Cnidaria is listed in the taxonomic coverage, but not as a keyword, or in the abstract or title.

pbuttigieg commented 5 months ago

@sformel-usgs

Thanks for posting.

We are indeed considering this type, especially as a means to bridge OBIS with other systems that have taxon-centric data.

We'll use your issue to initiate the pattern development, and I think it should satisfy your use case.

In the meantime, if you're publishing metadata using the @Taxon type, ODIS and OIH will be able to harvest it and - with a little percolation and validation - expose it.

@pieterprovoost - for your interest

pbuttigieg commented 5 months ago

The current plan is that:

Datasets about taxa (data-centric representation)

Datasets (or other CreativeWorks) will list the taxa that they are about using the schema:about property.

If the CreativeWork isn't really about the taxon, but is tangentially related to it, the name (and - if you use a schema:DefinedTerm stanza - identifiers etc) can appear in the schema:keywords property.

As this array can include several thousand entries, an approach to use a higher level grouping (e.g. Family, Order) may be used by the publisher (e.g. OBIS, Pieter Provoost suggested this approach). An alternative would be to have an @id JSON-LD property in the about value space pointing to the PID/URL of another JSON-LD file that lists the (large) array of taxa that work is about.

Taxon type as first order object (Taxon-centric representation)

One could also publish a JSON-LD file that is typed as a schema:Taxon itself. This would use WoRMS or other authorities to supply name and identifier values.

This type would allow you to list Datasets or other CreativeWorks about that taxon using a property like subjectOf, which may need nesting in additionalProperty given schema.org constraints.

pbuttigieg commented 3 months ago

The Taxon pattern has been created in https://github.com/iodepo/odis-in/commit/d5aa70394fe9d2b160639744898e0253dbf511dd

It's quite sparse, so we're going to need our OBIS, GBIF, OBON, ML2030 and other biodiversity communities to help define a rich pattern for (meta)data exchange.