Adding objects that should be searchable by their PIDs

[jkunze]

It would be great to make sure the following "charismatic" objects are added to the iSamples collection such that they are searchable by their PID (ARK or IGSN). It would also be good to understand such objects might be missing from the collection in the first place. Examples:

• https://doi.org/10.58052/DIA0000YL
• https://doi.org/10.58052/IEGIL000C
• https://n2t.net/ark:65665/337856f1a655e4ad78b1ef10a16dfb6e3
• https://n2t.net/ark:28722/r2p24/vdm_19600211

[ekansa]

I agree. Perhaps the fulltext search should include searching PIDs. Ideally, one should also be able to search for ARKs (both the legacy "ark:/" and the new expression "ark:" ) , DOIs, and IGSNs with and without a full URL expression.

So "https://n2t.net/ark:28722/r2p24/vdm_19600211" and "ark:28722/r2p24/vdm_19600211" and "ark:/28722/r2p24/vdm_19600211" should all return the same record.

See also: https://arks.org/news/2025-07-28-documentation-change-from-classic-to-modern-arks/

[rdhyee]

Ran your four examples against the production data file just now — they split into two different problems:

Example	In collection?	Why you can't find it
IGSN:IEGIL000C ("18LCI-5", SESAR)	✅ already in	Search box matches descriptive text, not identifiers → pure search feature gap
ark:28722/r2p24/vdm_19600211 ("VdM 19600211")	🔜 incoming	Among the 67,187 OpenContext records in the sync being prepared (#272) — verified present in Eric's fresh export
doi:10.58052/DIA0000YL	❌ absent	Not in the April-2025 export our data derives from — needs a fresher export from that source
Smithsonian ark:65665/337856…	❌ absent	Same: Smithsonian data is still at the April-2025 vintage

So: (1) search-by-PID is a real feature to build, and @ekansa's normalization list is the right spec — ark:/ and ark: (per the modern-ARK change), bare vs resolver-URL forms, DOI vs https://doi.org/…, IGSN with/without prefix. Acceptance test: paste any reasonable expression of a PID, get the record. (2) Missing records from non-OC sources is a separate data-refresh question — we'll track it distinctly so it doesn't get lost behind the search work.

— 🤖 rbotyee; table from queries executed against production data; RY reviewed