The document provides an overview of the technical infrastructure for the 3TU Datacentrum Back-end. It describes the Fedora repository software used for the institutional repository, Python middleware using Django and coupling to Fedora via its REST API, and an XSLT-based Fedora front-end with dynamic queries. It also discusses two types of datasets ingested - archival submissions recorded as-is for reference, and enriched objects which may be converted to other formats and have selected metadata stored in the resource index. Current datasets in the repository are listed.
3. Fedora Repository Software
Investigation of use in past year
Institutional repository implemented
Not perfect, but most flexible of comparable
systems.
The DIY downside is paradoxically also a key
benefit.
4. Python middleware
• Django Application framework
• Agile software development
• Fedora coupling via REST HTTP API
• Also use SOLR for indexing
• Fall-through to Fedora provided services
always possible
5. Fedora Front-end
• XSLT based
• Dynamic queries based on Resource Index
• Multiple output formats possible
• Used CMA for behaviours
6. Ellips: Fedora object
Pijl: relatie (rdf)
Kleine rechthoek: tekstuele metadata
Grote rechthoek: datastream (anders dan DC of RELS-EXT)
Diagram by E Gramsbergen
7. Fedora Front-end
• XSLT based
• Dynamic queries based on Resource Index
• Multiple output formats possible
• Used CMA for behaviours
demo link en link
12. Dataset types
Enriched objects
Possible conversions to other formats
For example CSV to XML
Manageable chunks
Selected metadata as RDF
Stored in Resource Index