[Archivesspace_Users_Group] OAI configuration questions

[Neal, Rick]

Good morning,

I have upgraded our version of ArchiveSpace to 2.5.0.  Recently I got a request to "get an OAI port configured so that Virginia Heritage can run a scraper and get our content off of ArchivesSpace".  They included some good instructions (see below) to help me.

My problem is that my config.rb file doesn't have any references to OAI in it at all.  The location I am looking at is ../ archivesspace/config/ .  There is a config.rb file in there but it doesn't match.

Is OAI configured in some other file for version 2.5.0?  Also, is there documentation for setting up OAI with version 2.5.0?

Thanks,

Rick Neal
Library Systems and Applications Administrator
University of Richmond

The instructions I received are below...

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Here are the default settings in config-defaults. These should be added and modified in local config.rb file:

# OAI configuration options
AppConfig[:oai_repository_name] = 'ArchivesSpace OAI Provider'
AppConfig[:oai_proxy_url] = 'http://your-public-oai-url.example.com'
AppConfig[:oai_record_prefix] = 'oai:archivesspace'
AppConfig[:oai_admin_email] = 'mailto:admin at example.com'
# In addition to the sets based on level of description, you can define OAI Sets
# based on repository codes and/or sponsors as follows
#
# AppConfig[:oai_sets] = {
#   'repository_set' => {
#     :repo_codes => ['hello626'],
#     :description => "A set of one or more repositories",
#   },
#
#   'sponsor_set' => {
#     :sponsors => ['The_Sponsor'],
#     :description => "A set of one or more sponsors",
#   },
# }
AppConfig[:oai_ead_options] = {}
# alternate example:  AppConfig[:oai_ead_options] = { :include_daos => true, :use_numbered_c_tags => true }


The two *necessary* options to set are:

oai_proxy_url should be set to the public URL. If not proxied and you keep the default http://localhost:8082/ for oai_url, you would replace localhost with the actual external hostname. ( same procedure for public webapp config. )

oai_record_prefix is used to generate what should be globally unique identifiers, so "oai:" prefix should be followed by a domain name. example:  'oai:http://lib.school.edu'


But you should also set a provider name ( that field was misnamed oai_repository_name  ) and an email address to report problems.

Sets are optional. I set up separate sets for law and health in our server, but I don't need to use these for VH crawl, as I'm going to be using the @mainagency code for distinguishing publishing org. So they should also have set Agency Code in ArchivesSpace "Manage Repositories" fields.

oai_ead_options support was added in v2.5.0. If you want digital objects exported, set that to be true.
For consistency, you should probably use the same options that you typically use when exporting from staff interface.
You probably don't want :include_unpublished,  as OAI is a public endpoint.


Once it's configured, you should be able to go to the  oai_proxy_url /sample and test the links on that page.

I would also suggest checking out the OAI implementation guidelines and the site validator in links below.
 The validator may not pass the FULL suite of tests, but should pass the basic ones.
( If it passes the basic test, the other problems are issues with the OAI implementation in ArchivesSpace. )


+ [Open Archives Initiative](http://www.openarchives.org/pmh/)
+ [Protocol documentation](http://www.openarchives.org/OAI/openarchivesprotocol.html)
+ [implementation guidelines](http://www.openarchives.org/OAI/2.0/guidelines.htm)
+ [site validator](http://www.openarchives.org/Register/ValidateSite)


~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Rick Neal
Library Applications and Systems Administrator
Boatwright Memorial Library
University of Richmond, VA 23173

rneal at richmond.edu



-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lyralists.lyrasis.org/pipermail/archivesspace_users_group/attachments/20181019/3df5bfb1/attachment.html>