Ldio:RepositorySink or Ldio:RepositoryMaterialiser > GraphDB

coret commented 3 months ago

I'm trying to make a LDES client to synchronize LDES to GraphDB.

I started out with defining the output in the pipeline as Ldio:RepositorySink as documented on VSDS Linked Data Interactions > LDIO Outputs > Repository Sink but when I posted the pipeline to the LDIO workbench the following error showed:

Pipeline "triple-store-pipeline": Invalid LDIO component in "Ldio:RepositorySink". Please validate by consulting the documentation.

Based on VSDS Technical Documentation > Consuming Pipeline > Output components > Repository Materialization (which has a non-working reference guide link) I switched to Ldio:RepositoryMaterialiser as output. Now the pipeline is accepted by the LDIO workbench (using docker image: ldes/ldi-orchestrator:2.5.1).

But, then next error occurs, as the log shows a lot of the following:

ERROR - when='sendToMaterialiser', problem='org.eclipse.rdf4j.repository.RepositoryException: java.lang.IllegalStateException: Transaction URL has not been set'

I now start to wonder if the config for Ldio:RepositorySink at VSDS Linked Data Interactions > LDIO Outputs > Repository Sink it correct/complete and is the same as the config for Ldio:RepositoryMaterialiser?

Below is my pipeline definition which generates the error above (the output config points to a GraphDB v10.6.1 instance on my local network which supports RDF4J v4.3.9):

name: triple-store-pipeline
description: "Requests all existing members from the LDES server and sends them to GraphDB"
input:
  name: Ldio:LdesClient
  config:
    urls: 
      - https://www.goudatijdmachine.nl/ldes/by-page
outputs:
  - name: Ldio:RepositoryMaterialiser 
    config:
      sparql-host: http://10.10.0.33:7200/
      repository-id: gtm
      named-graph: http://goudatijdmachine.nl/id/ldes-test

Yalz commented 3 months ago

Hi Bob, Thank you for the feedback. There apparently was a mixup with the documentation (the active develop documentation got published instead of the release). But as you figured out correctly , RepositoryMaterialiser is the correct name.

The configured pipeline should be correct. I'll try to reproduce your issue in the meanwhile. Do know that we'll be releasing a new updated version of this component soon (reduced memory usage)

Yalz commented 1 month ago

Documentation effort has been done and are now versioned

Informatievlaanderen / VSDS-Linked-Data-Interactions

Ldio:RepositorySink or Ldio:RepositoryMaterialiser > GraphDB #667