Versions
OverviewRecordsSchemasFiles
v1.1.0latest·51,093 records·0 files·70.2 MB
Aug 5, 20262
Collection
578 records
Community
153 records
Item
50,362 records
README

UMass Amherst ScholarWorks

The catalogue of UMass Amherst's institutional repository, held as records so other tools can read it without going back to DSpace.

Source

DSpace 7 at scholarworks.umass.edu. No authentication, no rate limit published.

  • Items over OAI-PMH (/server/oai/request), metadata prefix dim. The dim format keeps qualified fields that oai_dc flattens: dc.identifier.doi, dc.description.abstract, and the digcom.* fields left over from the migration off bepress Digital Commons.
  • Communities and collections over REST (/server/api/core/...), which reports parent links and UUIDs that OAI-PMH setSpecs leave out.

Record types

Type Count Notes
Item 50,362 One deposit. Item ids are handle suffixes, so item_22011 is 20.500.14394/22011.
Community 153 Departments, centres, programmes.
Collection 578 Items are deposited into these.

OAI reports 52,623 records and 2,261 of them are tombstones for withdrawn items, which are skipped. DSpace runs deletedRecord: transient, so those tombstones will not survive a future reindex. Half the harvest carries a DOI (25,305) and just under half an abstract (24,313).

Worth knowing

  • legacySeries is the department field. ScholarWorks has no department metadata. It does carry the bepress submission path (econ_faculty_pubs/53), and the segment before the slash names the old Digital Commons series, which maps to a department closely enough to filter on. Treat it as a hint, not an authority.
  • DOI coverage is partial and that matters downstream. umass/holdings joins this collection to OpenAlex on DOI first, so items without one fall back to title matching and score lower.
  • itemCount on containers is derived from the items actually harvested, not read from DSpace. It will disagree with the number DSpace shows if a deposit is embargoed or withdrawn.
  • Absent fields are absent. An item with no abstract has no abstract key rather than an empty one.

Run

pnpm scrape umass/scholarworks-items   # ~530 OAI pages, about 10 minutes
pnpm push umass/scholarworks-items --env dev