marc-portier
commented
5 months ago quite a number of aspects to consider:
- consider what the source/trigger of the logging actually is (meaning a running background service, like sembench vs ipynb stuff)
- actually producing sensible logging
- configuring docker-compose to actually direct all logging (now in the console) to files
- making those files accessible in the ipynb environment
E.g. when malformed queries in ./notebooks/queries are executed within a jupyter notebook they currently run but return no result. As an end-user, it would be handy if an error / warning is raised in the jupyter environment in this case.
exception raising for:
also about the access to logging in the different environments (where, topic, ...)