[QUESTION] Unable to run Ray example from documentation about fugue HOT 5 CLOSED

I see you closed the issue, are you good? This might be Ray version right? @goodwanghan

from fugue.

@juanitorduz i guess that is related with the ray versions. Ray changes behaviors in every version, so if you are using an old fugue version + a very new ray version, it could break.

Could you try

pip install fugue[ray]==0.8.6.dev4

and try again? It has the support of the latest ray. We will release 0.8.7 soon

from fugue.

Yeah! It seems something related to the versions. Our cluster runs with Ray 2.6.3 so I made sure now that I have the same version in jupyter hub (the latest Ray version was released yesterday).

I also made sure to pass fugue[ray] to the runtime_env argument of Ray init. I made this example work passing as_local=True .

I’m now grasping the fact that there a way from the examples from the docs to running this in a cluster. I'll keep you posted!

Thank you very much! 💪

from fugue.

I see you closed the issue, are you good? This might be Ray's version right? @goodwanghan

Indeed, seems it was an issue with the Ray version and the need to pass the packages (fugue[ray]) through runtime_env or Ray init.

I'm a Ray and Fugue nubie so I closed it because it seems something I could fix.

fyi: my end goal is to use statsforecast with Ray in jupyter hub.

from fugue.

I think I solved now with the right versions thanks! 🙏

from fugue.