[ywkim312]

Summary

• Give Organization, Person, Place, WebSite, Periodical, and Thing nodes a gleaner @id when they do not already have one: https://gleaner.io/xid/{type}/{host/path} from url, otherwise a name slug.
• Use an existing DOI (sameAs) or url for Dataset, ScholarlyArticle, and CreativeWork. Point a one-dataset folder's WebPage about/mainEntity at that dataset file's @id.
• Apply the same rule in scripts/generate_jsonld.py and document it in the JSON-LD prompt. Leave the ChatGPT sdPublisher node for AI: add (at)id to sdPublisher #20.

Test plan

• Confirm CHELSA webpage.jsonld publisher @id is https://gleaner.io/xid/organization/www.wsl.ch/en/index.html
• Confirm an existing @id (MRLC publisher https://www.mrlc.gov/#org) is unchanged
• Confirm a one-dataset WebPage about/mainEntity matches that dataset file's @id (for example G-RUN)
• Confirm a citation Dataset uses its DOI sameAs as @id
• Confirm ChatGPT sdPublisher still has no @id

[fils]

Actually, there is a best practice we should be following here..

https://example.org/.well-known/genid/

so for us

https://gleaner.io/.well-known/genid/

after genid/ we can do anything we want to something like

https://gleaner.io/.well-known/genid/xid/person/c-hari is fine, though xid seems redundant then.

ref: https://www.w3.org/TR/rdf12-concepts/#section-skolemization

[ywkim312]

Actually, there is a best practice we should be following here..

https://example.org/.well-known/genid/

so for us

https://gleaner.io/.well-known/genid/

after genid/ we can do anything we want to something like

https://gleaner.io/.well-known/genid/xid/person/c-hari is fine, though xid seems redundant then.

ref: https://www.w3.org/TR/rdf12-concepts/#section-skolemization

Good call. Updated the blank-node ids to Skolem IRIs and dropped the extra xid segment, so they look like:

https://gleaner.io/.well-known/genid/person/c-hari

https://gleaner.io/.well-known/genid/organization/www.wsl.ch/en/index.html

The WebPage and DataCatalog ids from #19 are still https://gleaner.io/xid/webpage/... and https://gleaner.io/xid/datacatalog/.... Those are real record ids, not skolemized blank nodes.

[valentinedwv]

remembering xid was supposed to be a repository code so some future URL resolver could resolve a uri: aka eco or earthcube

The WebPage and DataCatalog ids from #19 are still https://gleaner.io/xid/webpage/... and https://gleaner.io/xid/datacatalog/.... Those are real record ids, not skolemized blank nodes.

[ywkim312]

You're right. I treated xid as a fixed path, but that slot is the repository code a resolver would use later (eco, earthcube, and so on).

The CHELSA WebPage and DataCatalog ids still use the literal /xid/:

https://gleaner.io/xid/webpage/www.chelsa-climate.org/
https://gleaner.io/xid/datacatalog/www.chelsa-climate.org/datasets

Which repository code should this collection use in that slot, eco or earthcube? I'll switch those once that's set. The blank-node ids stay under https://gleaner.io/.well-known/genid/.

[valentinedwv]

lets go with earthcube, no use using the EarthCube Office (eco)

You're right. I treated xid as a fixed path, but that slot is the repository code a resolver would use later (eco, earthcube, and so on).

The CHELSA WebPage and DataCatalog ids still use the literal /xid/:

https://gleaner.io/xid/webpage/www.chelsa-climate.org/ https://gleaner.io/xid/datacatalog/www.chelsa-climate.org/datasets

Which repository code should this collection use in that slot, eco or earthcube? I'll switch those once that's set. The blank-node ids stay under https://gleaner.io/.well-known/genid/.

[ywkim312]

lets go with earthcube, no use using the EarthCube Office (eco)

You're right. I treated xid as a fixed path, but that slot is the repository code a resolver would use later (eco, earthcube, and so on).
The CHELSA WebPage and DataCatalog ids still use the literal /xid/:
https://gleaner.io/xid/webpage/www.chelsa-climate.org/ https://gleaner.io/xid/datacatalog/www.chelsa-climate.org/datasets
Which repository code should this collection use in that slot, eco or earthcube? I'll switch those once that's set. The blank-node ids stay under https://gleaner.io/.well-known/genid/.

Switched the CHELSA WebPage and DataCatalog ids to the earthcube repository code, not eco:

https://gleaner.io/earthcube/webpage/www.chelsa-climate.org/
https://gleaner.io/earthcube/datacatalog/www.chelsa-climate.org/datasets

The WebPage about and mainEntity point at that same DataCatalog id. Blank-node ids stay under https://gleaner.io/.well-known/genid/.

[valentinedwv]

looks good.