initial NCEI dataset JSON-LD template

[jmckenna]

• related to connect NCEI catalogue as an ODIS node #361

[pbuttigieg]

We passed an example dataset landing page with embedded JSON-LD/schema.org through the schema.org validator.

Here's the full address: https://www.ncei.noaa.gov/access/metadata/landing-page/bin/iso?id=gov.noaa.nodc:276264

We spotted a few issues to fix/improve

Things to improve/fix:

• keywords: this should be a JSON array rather than a single list with comma separation - the current form will be understood as a single keyword, which will reduce interoperability
  • the solution seems to be to correct the script generating the JSON-LD from the local metadata to generate an array
• @person and @organization values are NOAA generated URLs. These are likely not the IDs that the people and institutions would like to be identified with: ORCIDs, OceanExpert IDs, RORs, or other identifiers are to be preferred. Recommendation: NOAA should ask those that register to upload datasets to NCEI to enter their authoritative URL/PIDs for themselves and their institutions, and those should be shared in the metadata exposed to the web.
• GCMD keyword URLs are not resolving - something to check with NASA GCMD. Recommendation: for spatialCoverage values that are keywords or terms from a vocabulary, consider using DefinedTerm.
• The publishing Organization (NCEI) and the subOrganization (NOAA) listed under it may be inverted - this should be parentOrganization which is more sensible.
• General recommendation: the rich metadata for organizations and publishers in each record is very good, but may be more efficiently handled by creating and @Organization JSON-LD document with its own PID and linking off to it. It would be wise to do this only when there's a commitment to maintain these documents in perpetuity. For the moment, it's probably best to keep the repeated metadata in each @Dataset record.
• the distribution property seems to be misapplied: one would include URLs to directly download the dataset itself, along with metadata on the type of file, size, etc. Right now they take one to a landing page, which is somewhat redundant with the url property. These links exist, and look like this: https://www.ncei.noaa.gov/archive/archive-management-system/OAS/bin/prd/jquery/download/276264.1.1.tar.gz

[pbuttigieg]

Xref #361 (comment)

[jmckenna]

Closing, moved this to the new repo for templates instead: iodepo/odis-in#18