Oral Archivist Omeka S module

Releases

1.6.3

Oct 20, 2026

  • Reviewers, authors and researchers who open an Oral Archivist screen (the review queue, batch processing, an interview's review screen), and editors who open its settings, now get Omeka's Forbidden page with HTTP status 403; the review screen's and the settings' scripts get a JSON 403. They used to get Omeka's error page Omeka S encountered an error with status 500. Apart from the settings (below), who may use each screen is unchanged.
  • The Oral History tab on media pages showed every user its buttons (Process with Oral History AI, Import transcript JSON, View transcript, Retry processing), which led roles below Editor to that error page. The tab now shows them the status only.
  • Supervisors can no longer open the Oral Archivist settings, which hold the provider API keys, the budgets and the prompts. The documentation always reserved them to Global Administrators, but Omeka grants Supervisors everything it is not told to refuse them. They no longer see the Settings menu entry and get the Forbidden page. Give the Global Administrator role to anyone who manages providers.
  • The review queue's Settings button is shown only to Global Administrators. Editors saw it and were refused the page.

1.6.2

Oct 20, 2026

  • On the review screen, the interview's title and status now sit on their own line below the page header. At 1280 pixels wide a long title ran under the Save draft, Approve & publish and other buttons.
  • Chapter titles on the review screen get a full line of the Chapters panel, with the times and the move and delete buttons below them. They used to show about 15 characters ("Born above the b…").
  • The note above the transcript said "4 low-confidence passages flagged" while counting flagged words (each flag is one word). It now says "4 low-confidence words flagged".
  • Durations read the same everywhere: the review queue, batch processing and the media page's Oral History section now show 0:46 and 1:02:09, like the review screen and the public player, instead of 00:46 and 00:00:46.
  • In the Document layout, the search box above the transcript is wide enough for its "Find a word or phrase…" placeholder, which was cut off.

1.6.1

Oct 11, 2026

  • The Oral Archivist entry in the admin menu showed its text in a serif fallback font: the link carried Font Awesome's fa class, which sets the whole link in the icon font. It now carries only the icon class, like the other menu entries.

1.6.0

Sep 22, 2026

  • Page layouts: classic (default), sidecar (player + chapters left, transcript alongside), and document (reading view with inline chapter headings and a sticky player). Global setting with per-site override.
  • Placement setting (global + per-site): render the automatic public display before or after the page content.
  • AI summary: new summarization pipeline stage (skipped and free unless enabled), editable on the review screen, written to a configurable metadata property on the parent item at publish and removed at unpublish. New summary column and summarizing status. A Generate/Regenerate button on the review screen runs just the summary pass for interviews processed before the feature was enabled.
  • The pipeline now refuses to reprocess published transcripts.
  • Multi-media items: the item's summary property is now rebuilt from all currently-published sibling transcripts (one value per interview, in media order) — publishing or unpublishing one interview never clobbers another's summary. Theme media embeds AND media-list entries are hidden for media represented by the oral history viewer (wrapper-level, per media). Share/cite links are media-scoped (#t=125-m73) so deep links on multi-interview items target the right player; plain #t= links are honored by the first player only. Topic-page group headings append the media title when an item carries several interviews.
  • CRITICAL FIX: publishing no longer wipes the resource's other metadata. Partial API updates with collectionAction: replace remove every value not in the payload (core ValueHydrator behavior); both the extracted-text write (media) and the summary write (item) now use clear_property_values + append, which replaces only the target property. Installs that published with the broken versions should check affected items/media for missing values (titles especially).
Log in for Support