Depending on the analyst or user, different functionality is required:

• Researcher: Good Web UI for data entry for researchers, to maintain and update
• Analyst: APIs for system integrations and automatable workflows
• Analyst: Toolbox / utilities for using and combining data in analysis
• ITA: Reliable components that can be adapted to workflows at KTH

Current DiVA - future Invenio … probably?

The current DiVA repository have some strength and present some challenges:

• Used nationally in many institutions
• API lacking, slow progress towards CORA (DiVA API v3)
• Difficult to extend for KTH specific purposes

Invenio have some strengths in areas where DiVA does not:

• Invenio is similar to Zenodo, but more flexible, so known “globally”
• Good API, documentation
• Global community, actively maintained
• Extensible with custom fields
• Support for links / minted DOIs
• Presentation / GUI can be customized (can we make it look like DiVA, do we even want this?)
• Control over the data and the system, more empowered to quickly make adjustments
• Adapted by other institutions
• Comply to EU-standards for open data, a regional/global community

• Vad är poängen med dataflöden?
  • transparens & tillgänglighet
  • reproducerbarhet
  • analytics
  • standardisering

• Infrastruktur
  • Datatillgänglighet
  • Mellanlager för data (S3)
  • Knyta samman datakällor

## Andra tillgängliga datakällor och verktyg

• Migrate a few different records from DiVA to Invenio@KTH

  • Upload: Thesis, article, conference proceeding
  • Evaluate what is missing (for a specific usage / use case)?
  • Field mappings - Nazar has made a mapping for a complete view (including optional fields)
  • We could here first focus on required fields in DiVA
  • How to deal with missing values - a value is missing in DiVA, how will it be filled?

• Pre-populating or “seeding” Invenio with “vocabularies” requires YAML for
  • Organizations (ROR and internal KTH orgs)
  • Names/People/Authors (creators)
  • Resource types
  • Funding
  • Subjects (FOS, CESSDA - controlled vocabulary consortium och european social sciences, MESH) - namespaced with a specific - SwePub requires UKÄ-classification (at least UKÄ, Scopus, WoS etc)
  • Users

• Q: You can add, but you can not delete…. So a difference….
• Q: Are all “funders” pre-seeded/loaded - right now Swecris funders - ROR provides list of funders… and we can add Swecris and … What are the DiVA funders….
• Q: How do we match DiVA keywords to controlled vocabularies?
• Q: For Swedish keywords UKÄ and possibly an European one
• Q: Is name variations possible to load? Investigate.
• Q: How would it be possible to use different UI views for different resource types for a) data entry b) curation.

Description of the data flow / syncing process:

• We need to create 6 yaml files to load into Invenio (“vocabularies”)
• We then can use the API to “batch load” and sync data from DiVA
• The UI could be customized to support “custom fields”, see https://inveniordm.docs.cern.ch/customize/metadata/custom_fields/records/

Step 1: Some YAML and CSV files to migrate

https://gist.github.com/mskyttner/8b4e559a21229e494585f0fdc7119868/edit

Step 2: Create JSON and upload via the API

…

• Investigate existing customizations, such as https://infoscience.epfl.ch/