Sharing via OAI-PMH: Tips for Partners

Introduction

OAI-PMH is a protocol for harvesting the metadata descriptions of objects from an archive. It is the most common way that Plains to Peaks Collective (PPC) partners share their metadata with us, though we also accept metadata shared via spreadsheet, CSV file, and API.

In this guide you will find tips for making the OAI-PMH sharing process smoother between your institution and the PPC. While we have included links to OAI-PMH resources for some of the more common digital repositories (below), we are not experts in how to configure your specific OAI-PMH settings. Each system is different, so you will need to consult your user manual for information about making configuration changes.

OAI-PMH resources for common digital repositories

Islandora
https://github.com/Islandora/islandora_oai

Luna
https://doc.lunaimaging.com/display/V73D/OAI-PMH+in+LUNA

Contentdm
https://www.oclc.org/support/services/contentdm/help/server-admin-help/oai-support.en.html

Metadata schemes

Most institutions create metadata that complies with one of several possible metadata schemes. MODS, Dublin Core (DC), and Qualified Dublin Core (QDC) are the most common schemes that our partners use, though your institution might use a different scheme like METS or MARC, or even a locally created scheme.

Enabling your metadata scheme

It is important that your OAI-PMH feed is configured to share your institution’s optimal metadata scheme. For example, if your institution uses MODS, then you need to enable MODS in your OAI-PMH configuration. Otherwise your OAI-PMH feed will default to Dublin Core and your metadata won’t come through as granular or specific as you intended.

Tip: Be sure to double-check your OAI-PMH feed to make sure that it is outputting your metadata in the scheme you want.

Collections

The PPC harvests metadata at the collection level. We do not have the ability to harvest individual items or to omit specific items from a collection. In other words, we can’t cherry pick items from your collections.

Prior to each quarterly ingest, you will send us a list of collections (also called sets) that you want us to harvest. Ideally, you will also send us the collection name and the number of objects we can expect to find in each collection.

Here is an example of what to send:

Other tips

Turn off compound objects

You will need to disable compound objects in your OAI-PMH feed. For example, you might have a book cataloged as a compound object in your collection, which results in a separate metadata record for each page. By turning off compound objects, you will only be sending one record for the book as a whole.

Double check your OAI-PMH feed

Finally, always verify that you are actually sharing all of the fields that you intend. The PPC can only work with the fields that appear in your OAI-PMH feed, so if a field is missing or not mapped correctly it will not show up in the final aggregated metadata.