explicit identifier for fin import

[qjhart]

The current methodology for the fin import cli is;

• when an object has no identifier, and is a ArchivalGroup, then look to the schema:identifier and choose and ark: if it exists. For example the following would use ark:/87287/d7002v as the identifier

{
  "@context":  "http://schema.library.ucdavis.edu/context/dams.json",
  "@id": "",
  "@type": [ "fedora:ArchivalGroup", "CreativeWork", "Menu" ],
  "identifier": [ "ark:/87287/d7002v" ]
}

However, in practice, this would actually add the data as if it were explicitly set with the `"@id": "info:fedora/item/ark:/87287/d7002v"'.

If you have two identifiers that are arks: eg.

{
  "@context:": {
     "@context":  "http://schema.library.ucdavis.edu/context/dams.json",
     "shelfmark":"ark:/87287/d7r94j/"
  },
  "@id": "",
  "@type": [ "fedora:ArchivalGroup", "CreativeWork", "Menu"],
  "identifier": [ "ark:/87287/d7002v", "shelfmark:D-060/box:60+folder:112"]
}

Then it's unclear where to upload this data. The current plan is you can specifically add the @id parameter, and disambiguate this setup. This has two problems. One, we have to remember to add the info:fedora/item prefix, and second, in reality it's actually now ok in fedora 6 to have two @ids with valid LDP urls like:

{
  "@context:": {
     "@context":  "http://schema.library.ucdavis.edu/context/dams.json",
     "shelfmark":"ark:/87287/d7r94j/"
  },
 "@graph":[ {
  "@id": "info:fedora/ark:/87287/d7002v",
  "@type": [ "fedora:ArchivalGroup", "CreativeWork", "Menu"],
  "identifier": [ "ark:/87287/d7002v", "shelfmark:D-060/box:60+folder:112"]
  },
  {
    "@id":"info:fedora/ark:/9999/9999",
   "desc": "random info"
  }]
}

We are already using metadata with multiple @ids.

[qjhart]

I can think of a number of alternatives, originally, I was thinking that we use our own prefix like info:fin, and then make items like:

{
  "@context:": {
     "@context":  "http://schema.library.ucdavis.edu/context/dams.json",
     "shelfmark":"ark:/87287/d7r94j/"
  },
 "@graph":[ {
  "@id": "info:fin/ark:/87287/d7002v",
  "@type": [ "fedora:ArchivalGroup", "CreativeWork", "Menu"],
  "identifier": [ "ark:/87287/d7002v", "shelfmark:D-060/box:60+folder:112"]
  },
  {
    "@id":"info:fedora/ark:/9999/9999",
   "desc": "random info"
  }]
}

And then fail if there is more than one info:fin. But that wouldn't really survive an import/export/import,so I think the better alternative is to only use the @id if it's a ArchivalGroup and fail if there are two.

Alternatively, I suppose we could have a type like fin:Item and use that @id. Which is some way is better since you wouldn't need to scrub the fedora:ArchiveGroup and move it to the header?

[jrmerz]

Reminder, we have moved the info:fedora/item definition to be a @type based lookup in fin io import. The @type to base path (e.g. info:fedora/item) definition is stored at /fcrepo/rest/finio/config.json