OAI-PMH settings for Digital Collections

Digital Collections (DC) is built on Samvera Hyrax and supports the Open Archives Initiative
Protocol for Metadata Harvesting (OAI-PMH). This page reflects DC’s specific configuration of OAI-PMH. To find more about Samvera’s support for OAI-PMH, refer to its wiki.

DC currently only has one metadata prefix (oai_dc) that maps our robust DC metadata to simple Dublin Core for harvest by the Ohio Digital Network for the Digital Public Library of America (DPLA). See wiki page for oai_dc metadata mapping. Currently, only Rare Books &. Manuscripts Library’s (RBML) DC records are harvested by DPLA (link to DPLA records for RBML's DC records).

Base URL

The base URL for DC’s OAI feed is: https://library.osu.edu/dc/api/oai

The base URL leads to an error code of “badVerb”. However, the OAI feed is otherwise unaffected.

Screen grab of base URL with error code of badVerb.

Screen shot of base URL with error code of badVerb.

Internal identifiers

Within the system, administrative sets, collection objects, records (Generic Works), and files have internal identifiers. These identifiers can be used with the OAI-PMH feed with a caveat. Collection objects have not yet been configured with DC’s OAI feed. Record’s internal ID is colloquially referred to as a Generic ID. Admin sets have internal IDs, as well. Contact the Metadata Initiatives Librarian for the administrative set’s internal ID if needed.

Administrative Set

Works

Records are known as Generic Works, e.g. “George Bellows Record Book A”. At the end of the record URL (https://library.osu.edu/dc/concern/generic_works/xg94hr96b), is an alphanumeric code which represents the work’s internal identifier. Known as the Generic ID, it is after the https://library.osu.edu/dc/concern/generic_works/ string in the URL. Consequently, this record’s Generic ID is xg94hr96b.

Commands for DC’s OAI Feed

When entering commands for specific records and administrative sets, it is important to locate the appropriate system identifiers.

Explanation of commands

Commands are combined after the base URL to pull the selected metadata. However, this is an explanation of specific commands on their own.

Purpose	Command	Note(s)

Purpose	Command	Note(s)
Select metadata prefix	metadataPrefix=oai_dc	There’s currently only one metadata prefix so it is supplied here. However, there are plans to add more in the future.
List metadata formats available for harvesting	verb=ListMetadataFormats	See note above. This just shows the oai_dc metadata format.
Pulling one record	verb=GetRecord	Requires Generic ID
Pulling all records (first batch)	verb=ListRecords	Requires token from bottom of the feed page to get more.
Selecting a record or administrative set by ID	identifier=oai:library.osu.edu:[internal ID #]	To be combined with other commands.
List administrative sets	verb=ListSets	Currently, only administrative sets are configured with DC’s OAI feed. Collection objects are not.
Select to view more results when they exceed the first batch provided	resumptionToken=[Take token from bottom of of OAI output]	When done manually, token can be copied from the bottom of the screen to the end of the current page’s URL.

Command examples

This section shows examples of commands based on specific use cases.

Use case	Example	Notes

Use case	Example	Notes
Retrieving all records in DC	https://library.osu.edu/dc/api/oai?verb=ListRecords	This command would just be for the first batch. As the records span multiple pages, click the resume button at the bottom of the page or use the resumption token to continue retrieving results.
Retrieving a single record	https://library.osu.edu/dc/api/oai?verb=GetRecord&metadataPrefix=oai_dc&identifier=oai:library.osu.edu:xg94hr96b	In this example, the record example above, “George Bellows Record Book A,” is retrieved.
Pulling all records by an administrative set	Records by admin set (example), https://library.osu.edu/dc/api/oai?verb=ListRecords&metadataPrefix=oai_dc&set=unit:h702q854v	In this example, all the records for RBML are retrieved. As the records span multiple pages, click the resume button at the bottom of the page or use the resumption token to continue retrieving results.