Introducing the Texas A&M University Libraries Digital Asset Management Ecosystem




Creel, James
Bolton, Michael
Potvin, Sarah
Huff, Jeremy
Savell, Jason
Welling, William
Laddusaw, Ryan
Day, Kevin
Hahn, Douglas
Cooper, Micah

Journal Title

Journal ISSN

Volume Title



After several years of planning and technical development across Texas A&M University departments, the University Libraries are excited to announce the deployment of the first round of production-level services and applications comprising our Digital Asset Management Ecosystem. In this presentation, we will give a grand tour of the existing services and discuss our next steps. Our approach has emphasized a service-oriented architecture with separation of concerns between components and standard protocols for information transfer. This has enabled us to integrate legacy components into the same workflows as new ones. In particular, our legacy DSpace instance, OAKTrust, participates on a par with a new Fedora repository, and both repositories can receive content from our ingestion tools and use that content to drive user-facing discovery and exhibition layers. Conduits for curation and ingestion of content include legacy workflows with DSpace SAF (Simple Archive Format), SWORD (Simple Webservice Offering Repository Deposit) from Vireo, and various command-line scripts. New, more user-friendly workflows use RESTful APIs through the MAGPIE (Metadata Assignment GUI Providing Ingest and Export) application that has been presented previously at TCDL. The MAGPIE application can bring in metadata from our Voyager catalog, CSV spreadsheets, DSpace SAF exports, and automated suggestions from controlled vocabularies. The content (PDF or image) and metadata are then displayed in the system for a human to edit and amend. Publication over REST APIs is currently available for DSpace, Fedora, and Archivematica. MAGPIE can also operate in a “headless” mode if no human curation is required. In “headless” mode ingested content is published immediately to the destination. Content available in our DSpace and Fedora IRs is of course exposed via the out-of-the-box interfaces these systems provide. For DSpace, these interfaces include the XMLUI, Solr, and an RDF webapp. For Fedora, these include Solr, Fuseki, and a robust messaging service. In addition, Fedora now offers a facility called API-X for proxying and modifying HTTP requests to Fedora in interesting customizable ways. One important development in this framework is the PCDM extension from Amherst College, which provides RDF metadata for PCDM-structured objects in your Fedora repository. We use this extension to drive a new IIIF manifest generator that generates Collection or Presentation manifests compatible with a variety of services, including Spotlight, Mirador and the Bodleian Libraries IIIF Manifest editor. In the future, we plan to enhance our IIIF manifest generator to utilize RDF responses from the DSpace RDF webapp in the same way it does from the Amherst PCDM Fedora extension. We will also continue to deploy new user-interfaces for discovery and exhibition. In this regard, we are pleased to have the flexibility to use custom in-house solutions or existing open-source projects, so long as they adhere to standards such as well-defined REST APIs, PCDM-RDF, and IIIF.