-
-
Save greglandrum/d721822e8973ff8438cf2af283c9271e to your computer and use it in GitHub Desktop.
@maxime-beck : because this needs to launch a KNIME Analytics Platform instance each time you run a workflow, it's always going to take a while. I find it most useful when combined with workflows that already take a long time to execute.
If you're a KNIME Server customer: there will be an updated version of the package soon that can take advantage of a KNIME Server to execute the workflows. This will allow a much lower latency than we get when using things locally.
@greglandrum I see that the library is invoking the KNIME Batch Application but that CLI doesn't provide a way to just start an instance and keep it running. Is there any plan to add that feature to the Batch Application anytime soon? That would greatly improve the performances of the Jupyter integration.
@mbezy: I have the same question. Have you found the answer yet?
Hi, thanks for sharing the code!
This is very very useful.
This is very useful thank you!
In terms of execution time, would there be any optimization possible? It's quite long to execute even the smallest workflow (8 seconds to create and sort 4 cells). I would assume most of this time is used to load the workflow?