Florian M
Florian M
Looks like no trees in the NML is interpreted as no skeleton layer should exist. We should probably add a property to NML files to indicate an (empty) skeleton layer
I’m sure scala/java has support for asynchronously reading (byte ranges of) local files. That should be used in the FileSystemDataVault. This should free more threads
dataset uploads are a critical feature. We should discuss how we can add end-to-end tests for it, and which cases to cover.
Follow-up for https://github.com/scalableminds/webknossos/pull/7386 (We didn’t want to change the worker in that PR as it is already pretty big) Should be adapted both in JobController and AiModelController (and of course...
Bucket requests always return 200, if something went wrong during loading, the array is all-black and the MISSING-BUCKETS header is set. However, no error information is contained. There is already...
In an annotation with a volume annotation layer, the reload button will call the datastore triggers/reload route, but with the tracingId as layer name. The datastore does not know the...
Hundreds of functions in datastore + tracingstore need to pass around the user token for authentication. This could be an implicit parameter, just like the DBAccessContext. Would need to be...
When executing long-running jobs, the user can currently only see the state as one of Pending, Running, Success, Failed. It would be cool if the user could get access to...
We should determine the size of the to-uploaded files already in the frontend and check if they exceed the quota. reserveUpload should send either a total file size or even...
Backend logging debug level is not shown in production, revisit code, decide what should be promoted
We should either change the config to log debug to stdout in production, or decide which logging should be promoted to info/warning/error.