Group (near) identical records

[pbuttigieg]

We expect that partners will host copies of digital assets, leading to multiple identical or near-identical results. On the UI and in the UX, these should not appear as independent results but - collectively - in a grouped result. That is, one result element with a visual sub-element stating that "[0-9] versions available". Clicking on the sub-element

• [ ] IODE to decide whether the identification of (near-)identical records occurs in the ODIS Graph or in the OIH SOLR index.
  • [ ] If ODIS-level detection is chosen, update Dashboard to report on degree of redundancy.
• [ ] Determine which record should be the "displayed" record (metadata completeness?)
• [ ] implement grouping on the schema:sameAs property
• [ ] define grouping rules based on other keys (e.g. name, description, identifier, ...)
• [ ] implement UI/UX elements to handle grouping and display/expand groups on click, noting different providers.

[jmckenna]

The RDA partner's interface handles these duplicates through a small arrow/dropdown near the top, see "3 linked records" at : https://researchdata.edu.au/global-fisheries-landings-archived-version/794636