Comet Sprint 53 / Daylight Sprint 25 Review (Apr. 29, 2026–May 18, 2026)

Spring Workcycle: Apr. 29, 2026–May 18, 2026

What was the goal of the Comet Sprint 53 / Daylight Sprint 25 (Spring 2026, Sprint 1)? 

Advance discovery integration (so we deliver a connected system at the end of the workcycle) and unblock production data migrations.

What is the milestone that this sprint is supporting? 

1. Discovery Integration 
   1. Publishing/unpublishing from Comet to the discovery prototype
   2. Preparing for UCSD discovery development/integration
2. Discovery Development
   1. Exploration into advancing front-end development
3. Interim solution for controlled values
   1. Architectural/infrastructure/design work
   2. Implementation into Comet and Goldengate
4. Comet Ingest Improvements & Reliability

Accomplishments of Comet Sprint 53 / Daylight Sprint 25

1. Discovery Integration (David M., Margaret)
   1. Publishing and unpublishing collections, objects, and files from Comet to Golden Gate
   2. Golden Gate API support for file downloads
   3. Golden Gate API support for using ARKs as identifiers (partial UI support)
   4. Cleaning up Superskunk API endpoints to make them more maintainable
2. Discovery Development (Alissa/David T.)
   1.  Began refactoring to become production-ready.
3. Interim solution for controlled values (Margaret)
   1. We have come up with a general approach for storing controlled values in the repository in CSV or TSV format and being able to load them into both Comet and discovery platforms
   2. Work on getting this fully integrated in Comet is ongoing—the main difficulty has been in getting the edit page to support the very long lists of values that result
   3. Work on getting this integrated in Goldengate is still pending
4. Comet Ingest Improvements & Reliability (David M., Michael S., Matt C.)
   1. Added ARKs to collections display page
   2. FITS xml configuration changes
   3. Thumbnail generation and reliability (investigation)
      1. PDFs & Pyramidal TIFFs—potential race conditions even after a few bug fixes that result in some small number of thumbnails missing; it’s also random across repeat runs of the same importer
   4. Large file size ingests and leaked temp files (investigation)

What’s next?

• Discovery integration
  • Further API improvements:
    • respect visibility flag
  • Further search improvements:
    • keyword search against properties
    • search/filter by date range
    • sorting support
  • Address Golden Gate / Daylight development issues around sample data and campus metadata schemas
• Discovery development
  • Complete support for ARK-based URLs
  • Build UI for search by collections display
  • Continue work on refining Complex Object display
  • Continue production readiness refactors 
• Comet 
  • Finish up controlled value support & turn on for each campus.
  • Discuss alternate pathways/solutions for both thumbnails and large file/leaked temp file cleanup. Perhaps considerations for making derivatives outside of Comet but supported via Bulkrax for imports (e.g., for thumbnails). 
• Sending metadata to EZID and possibly DOI support. 

GitLab link: https://gitlab.com/groups/surfliner/-/boards/9240933?iteration_id=3870434