Skip to content

RFC 091: Digitisation ingest identifiers during the Sierra to Folio migration#159

Open
kenoir wants to merge 3 commits into
mainfrom
rk/rfc-091-digitisation-ingest-identifiers
Open

RFC 091: Digitisation ingest identifiers during the Sierra to Folio migration#159
kenoir wants to merge 3 commits into
mainfrom
rk/rfc-091-digitisation-ingest-identifiers

Conversation

@kenoir

@kenoir kenoir commented Jun 29, 2026

Copy link
Copy Markdown
Contributor

Preview

View the rendered RFC on this branch (renders the Mermaid diagrams and tables):


What does this change?

Publishes the digitisation-ingest discovery work as RFC 091. The RFC sets out how digitisation ingest identifiers should work across the Sierra to Folio LMS migration.

Today the Sierra b-number is the shared key for digitised content. It is the Goobi ingest package id, the storage externalIdentifier (digitised/<b-number>/vN), the METS work's own identity (mets/<b-number>), the merge candidate that joins the METS work onto the public catalogue work, and the IIIF manifest URI and DLCS asset ids. The migration removes the b-number, so the RFC explains why it cannot simply be swapped for a Folio id (doing so would reset storage versioning and silently break the merge join key, leaving digitised content invisible on the public record with no error), and proposes minting catalogue-style identifiers at ingest instead.

The proposal builds a small application for digital production staff that creates ingest package metadata from Folio, and splits behaviour on whether the Folio record carries a predecessor pointer to an old b-number:

  • Cross-migration (predecessor present): reuse the old b-number as the storage handle and keep emitting the sierra-system-number merge candidate. This is the lower-risk path; it needs no transformer change and leans on RFC 083 predecessor relationships to keep canonical ids stable. After migration it needs only Folio-target merger rules.
  • Post-migration (no predecessor): mint a catalogue-style id (C) via a secured resolve-or-mint endpoint added to the Identifiers API (RFC 089). C becomes the storage handle; the Folio id rides in the METS record identifier so the transformer can emit a folio-instance merge candidate and build the mets/<folio-id> own-identity. The transformer carries C from the storage externalIdentifier into the work's identifier block, and the id_minter verifies the mets/<folio-id> ↔ C binding (raising rather than minting on a mismatch). This path is blocked on RFC 081's IIdentityService in iiif-builder.

The RFC also covers alternatives considered, the merger-rule changes needed for Folio targets, dependencies for each path, and two open questions.

Files changed:

  • rfcs/091-digitisation-ingest-identifiers/README.md: the RFC document (with two Mermaid sequence diagrams for the current and proposed ingest flows).
  • rfcs/README.md: RFC listing table regenerated by .scripts/create_table_summary.py (RFC 091 row added).

How to test

  • Read the rendered RFC and review the context, proposal, alternatives, dependencies, and open questions.
  • Confirm the RFC passes repo validation: .scripts/validate_rfc.py.
  • Confirm the listing table is in sync: .scripts/create_table_summary.py --check-readme.
  • Some relative links (../085…/, ../088…/, ../089…/, ../090…/) resolve only once those branches merge; they are correct against the merged layout.

How can we measure success?

No measurable runtime success criteria; this is a documentation RFC. Success is the RFC being reviewed and the team aligning on the digitisation-ingest identifier approach across the migration, in particular the decision to mint catalogue-style ids rather than reuse the Folio id, the cross-migration / post-migration split, and the secured mint endpoint on RFC 089.

Have we considered potential risks?

  • Documentation only; no production code or infrastructure is changed by this PR, so there is no runtime or deployment risk.
  • The design risks are enumerated in the RFC and are the subject of review. The principal one is the silent merge break that follows from moving the package id off the b-number; the proposal mitigates it by reusing the b-number on the cross-migration path and by a fail-loud guard on the post-migration path (the id_minter raises rather than minting a divergent id when a METS source identifier is not already bound to its storage handle).
  • Two open questions remain, both flagged as gating the post-migration path: confirming iiif-builder and DLCS make no assumption that the storage externalIdentifier (C) equals the METS record identifier (folio-id); and confirming the id_minter can verify a supplied source-id ↔ canonical-id binding and raise on a mismatch, rather than only assigning a fresh id.

Notes for reviewers

🤖 Generated with Claude Code

kenoir and others added 3 commits June 29, 2026 09:48
…igration

Publishes the digitisation-ingest discovery work as an RFC. It sets out
why the Sierra b-number cannot simply be swapped for a Folio id during the
LMS migration, and proposes minting catalogue-style identifiers at ingest
via a secured endpoint on the Identifiers API (RFC 089), with separate
cross-migration and post-migration ingest paths.

Regenerates the RFC listing table via .scripts/create_table_summary.py.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
…ring

Mermaid treats ';' as a statement separator in sequence diagrams, so the
verify-step message split mid-line and the diagram failed to render.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
…s and arrow symbols

Open question 1 now states the iiif-builder/DLCS concern in terms of the
storage handle vs the file names in METS, and points to RFC 085, which owns
the IIIF/DLCS identity rework and currently lists the Goobi message and METS
shape among its open unknowns. Also replaces em-dashes and → / ↔ symbols
with plain prose.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
@kenoir kenoir marked this pull request as ready for review June 29, 2026 09:37
@kenoir kenoir requested review from a team as code owners June 29, 2026 09:37
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant