[FEATURE] Extend OAI-PMH support to allow fully OAI and DINI compliant instances

[stweil]

Description

Problem

The current implementation misses some features which are either required bei OAI and/or DINI or suggested by DINI:

1. "Identify" does not support the optional element "description" which is suggested by DINI.
2. Set names with underscores or colons are currently not fully supported (see [BUG] Fix creation and validation for OAI-PMH setSpec values #1480). Meanwhile this is fixed.
3. Set names with colons need special code because they represent hierarchical sets.
4. We urgently need documentation how to add OAI-PMH to a TYPO3 instance with Kitodo.Presentation. See also [DOCUMENTATION] Update Documentation #1178.

Proposed Solution

1. It should be possible to add the required data for "description". Then this data should be used in the answer for an "Identify" request.
2. Set names must allow underscores or colons. This is now possible with the latest code.
3. Hierarchical sets must be fully supported.
4. Documentation which handles all required settings has to be written.

Mockups and Examples

Expected Benefits of this Development

Improving compatibility with the DINI requirements for open access publication services would allow better services.

Estimated Costs and Complexity

I expect that this requires medium resources.

Target Version

The new features should be added to the main branch. Backports to older releases are optional.

Additional Context

• https://www.openarchives.org/OAI/openarchivesprotocol.html#Identify
• https://edoc.hu-berlin.de/server/api/core/bitstreams/608901cd-de04-470d-9f34-9799e2570bdd/content (M.9-6)

[stweil]

I just noticed that Kitodo.Presentation does not support underscores and colons in the setSpec (oai_name). DINI requires such sets (for example "open_access", "doc-type:Periodical"), and they are well defined by OAI, so this might be a bug.

Update 2025-02-07: There is also an issue with uppercase characters. Details are given in issue #1480.

[sebastian-meyer]

Missing underscores in set names is definitely a bug.

But support for colons would be more complex than just allowing the character in set names. Colons in OAI sets are meant to represent hierarchy, so if you have a set named doc-type:Periodical and doc-type:Monograph the OAI specification requires the API to return both sets if someone requests a set doc-type see https://openarchives.org/OAI/openarchivesprotocol.html#Set).

[stweil]

That's correct. Ideally Kitodo.Presentation should fully support hierarchical set names. But allowing the colon would be a first and easy step which allows starting with such set names. Then a newer software release with extended code would automatically do the right things.

Should we extend this funding issue to add more complete OAI support?

[sebastian-meyer]

Should we extend this funding issue to add more complete OAI support?

I don't know, it's your proposal! ;)

[stweil]

Examples of set names for DINI: ddc:070:073, doc-type:Periodical, doc-type:Text, open_access, status-type:publishedVersion.

These names can now be entered, but the names with ":" represent hierarchical sets which need special handling in the code. For example ddc:070:073 is implicit also part of the sets ddc:070 and ddc.

[sebastian-meyer]

Votes: 1