JudaicaLink in production

JudaicaLink started as a research project around 2012. Since 2016, the development is funded by the German Research Foundation as part of the Specialized Information Service Jewish Studies (FID Jüdische Studien) at the University Library Johann Christian Senckenberg in Frankfurt am Main.

Why the move

While JudaicaLink originally was a stand-alone Linked Data project, it became clear that the knowledge graph ultimately should be integral part of the FID services and run in Frankfurt to ensure long-term availability and to facilitate further applications and developments built on top of JudaicaLink within the FID portal. JudaicaLink within the FID portal allows a closer integration with the subject catalogue, the repository and the authority data work of the FID.

Moving JudaicaLink into the FID portal was therefore part of the mission of the since third DFG funding phase (2023 - 2025). This phase actually saw two migrations: first, in 2023, JudaicaLink moved from Stuttgart Media University to Mannheim Technical University. Then, in 2025, the actual migration to Frankfurt happened in close coordination with the development of the FID portal.

What changed technically

The move was not a copy. The service was rebuilt to fit the portal:

• JudaicaLink is integrated as several Django applications in the FID portal rather than running as a separate system.
• The search backend was migrated from ElasticSearch to Apache SOLR already in Mannheim, ahead of the move, since SOLR is what all search applications in the Frankfurt portal use. All indices were rebuilt, and data structures and source code were adapted.
• The service was containerized, again already in Mannheim ahead of the move, to make deployment simpler and more flexible than the previous server setup allowed.
• Data ingestion was consolidated. The Django applications that control loading and indexing datasets were cleaned up and brought together, so that fewer, clearer components do the job.
• data.judaicalink.org now resolves to Frankfurt. Entity URIs keep working and serve HTML, RDF/XML, Turtle or JSON-LD through content negotiation, so references stored in other applications remain valid.
• Dumps and the BEACON file are delivered from Frankfurt. The German National Library receives its updates for EntityFacts from there.

Who does what now

Frankfurt runs the service. Operation, data curation, the dataset catalogue and user support are with the FID. The dataset documentation is maintained there, and that is where additions and corrections belong. For questions about the content of JudaicaLink or about the service itself, the FID team in Frankfurt are the people to ask.

Mannheim develops. This is where work continues on the knowledge graph itself: the data model and ontology, provenance of individual statements using RDF-star, the planned integration with Wikidata, and the processes that turn heterogeneous source data into Linked Open Data. What proves itself here moves into the portal.

The older systems

Three systems were in use during development. Only this website remains in Mannheim:

By the way, this was always the reason for the distribution of the JudaicaLink service across different subdomains and servers. We want to keep the entry point stable, but systems should be able to move.