Hi Everyone, We’re in the process of relaxing our ...
# questions
m
Hi Everyone, We’re in the process of relaxing our kedro dependencies in our repository. The current Kedro version is
==0.18.3
. When we relax it to
~=0.18.3
, pip would install
0.18.8
while compiling and we get the below error for some of our pipelines:
Copy code
KeyError: 'logging'
Can you provide some pointers what could be the reason behind this?
0.18.*
should have no breaking changes based on the RELEASE notes so I’m not sure what could explain this. Thanks for the help!
j
@Mate Scharnitzky all the 0.18.x releases should be backward compatible yes. are you able to share a full traceback to know where does that come from?
m
This is what the log looks like on CI.
Ok, actually I’m getting confusing info about what kedro version CI is using. Let me clarify that.
👍🏼 1
j
for the record, the
logging_config = self._get_config_loader()["logging"]
line was modified in 0.18.4 https://github.com/kedro-org/kedro/commit/997692a22da1a67f220b3c4cf8e3935076d28e11 not sure if related
could be something else
m
thanks! I’ll look into it
a
Are you using a custom configloader?
m
We use Jinja and we should deprecate that soon
I don’t think that’s the issue as one of the apps don’t fail and it also uses Jinja
Let me get back to you tomorrow with more thoughts from the team.
If there is nothing obvious from your side, then my hunch is that there might some issues how the dependencies are installed or some package config error.
a
I’m pretty sure we haven’t seen this problem before so indeed I think it’s something peculiar to your setup. I would double check what version of kedro is being run here.
It’s plausible there’s been something overlooked on our side with the changes to the configloader, but if that was the case then I’d expect you to get the same problem on other apps with the same configloader and for many other people to have similar problems.
Let us know how you get on though, because this is not a very revealing error message so I’ll be curious to see if it’s something we should improve anyway!
m
Thanks, Antony! I’ll keep you posted.
👍🏼 1
Hi Antony, Juan,
I just wanted to get back on the above topic.
We haven’t found the exact root cause of the problem, but we’re quite sure it’s in connection with our custom config loader + multi-runner.
We still use jinja templated config loader but it’s in our backlog to investigate how config should be done with omegaconf + multi-runner, so we will upgrade our kedro version as part of that effort.