-
Notifications
You must be signed in to change notification settings - Fork 1.6k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[Engprod] Stable python package management #3078
Comments
I think we have two goals here:
Python's way of pinning dependencies is to freeze the packages to a file: So, we should probably test both ways - test with pinned packages installed from requirements.txt and test with latest packages installed by running the setup.py. We should probably use pinned versions for the image building. |
@Ark-kun Thanks, totally make sense to me. Can we handle this goal in post submit test instead, so it is not presubmit blocking? |
Sounds good. |
TODOs:
|
|
Metadata writer image: #3408 |
The visualization-server licenses seem to be broken after some package got updated. |
@Ark-kun OK, I will try to do the same for visualization-server. |
This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions. |
/lifecycle frozen |
(kubeflow#4310) * build: fix visualization server build failure by adding missing licenses for new deps * wip * chore: pin visualization server python dependencies * updates * update licenses
(kubeflow#4310) * build: fix visualization server build failure by adding missing licenses for new deps * wip * chore: pin visualization server python dependencies * updates * update licenses
(kubeflow#4310) * build: fix visualization server build failure by adding missing licenses for new deps * wip * chore: pin visualization server python dependencies * updates * update licenses
We have done many hotfixes like
Also e.g. things don't work when we upgrade base google/cloud-sdk image.
Can we set up a stable python package management strategy that pins all the packages we use?
Some options:
UPDATE: we have decided to go with pip-tools, because it's lightweight and just focus on the dependency pinning task we wanted.
/cc @numerology
/cc @Ark-kun
/area engprod
The text was updated successfully, but these errors were encountered: