Workflow for syncing UMIL labels with Wikidata

[PouyaMohseni]

As discussed, we need to update our data via Wikidata pulls. This requires a reconciliation workflow between the databases.

Since UMIL and Wikidata are asynchronous, several scenarios can occur.

Possible Changes

In UMIL (per instrument, per language)

• Names can be added (with different verification statuses)
• Names can be deleted (tracked via soft delete, [New Instrument Name] reassign umil_label after instrument name is (soft) deleted #350)
• Another name in the names list may become the UMIL_label

In Wikidata (per instrument, per language)

• The main label may be changed
• The main label may be removed
• Aliases may be removed
• Aliases may be added

Moreover,

• New UMIL names are checked at the time of addition to avoid duplication with Wikidata entries.
• Some UMIL names may also be batch-added to Wikidata.

Proposed Rules

• New alias or label appears in Wikidata

  • If it already exists in UMIL (soft-deleted or active) → do nothing
  • Otherwise → add it to UMIL

• A name in UMIL that was sourced from Wikidata is removed from Wikidata

  • → Hard-delete the name from UMIL*

• Wikidata main label differs from the instrument’s UMIL_label

  • If both labels are sourced from Wikidata → update UMIL to match Wikidata
  • Otherwise → do nothing

* Names refers to both labels and aliases in the UMIL context.

** This is a hard delete because the soft-delete feature exists to support UMIL’s internal review process, not Wikidata’s.
An instrument label may be removed from Wikidata and still require review if it originated there, but already rejected entries do not require review.