PubMan Func Spec Export/OAI Data Provider

From MPDLMediaWiki
Jump to navigation Jump to search

Description[edit]

A set is an optional construct for grouping items for the purpose of selective harvesting.

Scheduled[edit]

for PubMan R5

Definition[edit]

TODO:

  • what would be good criteria to define sets?
  • what kind of sets we would like to provide to outside?
  • to be able to check if the set definition can be fulfilled by filters
  • FIZ will check if it is able to expose filters as sets
  • MPDL will check better requirements and put on Colab


OAI-PMH Set Definition[edit]

eSciDoc

eSciDoc:Solution

eSciDoc:Solution:Context

As contexts are only administrative objects, we will not offer them for OAI-PMH, they even have no PID. In previous discussions we agreed that if we want to have sets equal to contexts, we will all objects to a Container. --Natasa 08:50, 16 December 2008 (UTC)

eSciDoc:Solution:OrgUnit

eSciDoc:PubMan:Yearbook --Natasa 08:50, 16 December 2008 (UTC)

--Natasa 09:16, 16 December 2008 (UTC)Sets can also be defined by subject category or other controlled vocabulary, such as:
  • eSciDoc:PubMan:physics
  • eSciDoc:PubMan:OpenAccess

Examples:

Requirement Set
Get me all PubMan items of my institute eSciDoc:PubMan:MPDL
Get me all YearBook PubMan items eSciDoc:PubMan:Yearbook
Get me all items from Collections MPI/PL Grey Material eSciDoc:PubMan:MPI/PL_Grey_Material
Get me all Pictures from FACES eSciDoc:FACES
Get me all PubMan items from Sengbusch Collection eSciDoc:PubMan:Sengbusch
As stated above, both Sengbusch and PL_Grey_Material shall have extra Containers if we would like to offer them as sets --Natasa 08:52, 16 December 2008 (UTC)

OAI-PMH Set Definitions for FACES[edit]

In general, for me it's not clear, for what we need an OAI interface for Faces, because only the metadata will be exported and I don't see any scenario where someone wants to reuse the metadata from Faces. But when sets shall be implemented, I would propose following sets (--Kristina 10:04, 23 December 2008 (UTC)):

  • eSciDoc:Faces:young
  • eSciDoc:Faces:middle-aged
  • eSciDoc:Faces:older
  • eSciDoc:Faces:male
  • eSciDoc:Faces:female
  • eSciDoc:Faces:picture set a
  • eSciDoc:Faces:picture set b
  • eSciDoc:Faces:anger
  • eSciDoc:Faces:disgust
  • eSciDoc:Faces:fear
  • eSciDoc:Faces:happiness
  • eSciDoc:Faces:neutrality
  • eSciDoc:Faces:sadness
I think you are right, that it probably makes no sense to have Faces exposed via OAI-PMH interface. However, maybe albums cam be exposed? --Natasa 10:15, 23 December 2008 (UTC)

General Ideas[edit]

  • The sets can be retrieved via filter function (Filter needed for Solution, orgUnits, Contexts)
    • Sets can be defined as "saved filters" with description and defined criteria
  • The search&export interface can be enhanced with export format 'oai-pmh'. Where the user can define his own oai-pmh set via the search function.
    • what is the idea behind having export format oai-pmh? The oai-pmh is separate interface. Via this interface we can offer repository content in several metadata formats. --Natasa 09:37, 16 December 2008 (UTC)
Agreed. I also would not go for that idea. --Nicole 17:12, 17 April 2009 (UTC)
  • The Data acquisition service provides fetching of the standard oai-pmh sets.
    • why would data acquisition service provide fetching of the oai-pmh sets? My understanding was that the data acquisition service provides fetching a single record from external/internal source and offering this very same record into another format. Actually, the question is why would data acquisition service had knowledge on oai-pmh sets? --Natasa 09:37, 16 December 2008 (UTC)
    • I think the OAI-PMH interface should be part of the dataAcquisition service as this is just another protocol of fetching escidoc items (now the dataacquisition service provides fetching escidoc records over unapi). in my opinion OAI-PMH should be the next protocol in the very same service.--Kleinfercher 10:19, 16 December 2008 (UTC)
      • actually, oai-pmh may be more generally applicable for automated metadata acquisiton, because it guarantees the existence of one metadata format, namely oai_dc. so yes, i too think the data acquistion service should know how to talk oai-pmh.--Robert 17:53, 16 December 2008 (UTC)
  • The set description is generated dynamically because it will contain all orgUnits and Contextes
    • hm. responding to the listSets verb may be a bit of a problem with the "sets as stored searches" approach.--Robert 17:56, 16 December 2008 (UTC)

Base URL[edit]

http://coreservice.mpdl.mpg.de/escidoc-oaiprovider

References[edit]