Closed ludovicm67 closed 2 years ago
That's probably because bnode IDs are preserved right now on the amos DB, so we create clashes if we re-use IDs. This can be fixed by taking the DB offline, change the preserve.bnode.id
setting to off & re-write the data (after a clear).
The setting is already off
, so this is not the issue.
Maybe the fault of some old data?
I will clear everything in that database and deploy everything back there by running the deploy
step from the pipeline.
It's now able to generate working manifests again.
We have a Trifid plugin who is in charge of building the right IIIF manifest. But we currently have some issue: they look broken.
When I investigated last time, the issue was coming from the way the triplestore was handling the blank nodes.
CLEAR
the graph, triggering the pipeline again and clearing the cache was solving the issue last time.Maybe using
PUT
instead ofPOST
when writing in the graphs may solve this issue?