Skip to content
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

CLN: Exception in io, plotting #28649

Closed
wants to merge 2 commits into from

Conversation

jbrockmendel
Copy link
Member

No description provided.

@jbrockmendel
Copy link
Member Author

@datapythonista the two failures in the slow build are showing up for me locally (OSX, py37, up-to-date numpy). I'm having a hard time troubleshooting them. Any idea who to talk to?

@datapythonista
Copy link
Member

Not sure if I understand. Do you mean you can't reproduce them locally?

My view is that we shouldn't be using pyplot. pyplot is a convenient API for users in a notebook. But for production code using an API that changes an internal state of a module like pyplot does makes things very tricky. Mainly for the tests, since tests running in parallel keep changing the internal state of pyplot affecting each other. Matplotlib offers more Pythonic APIs.

Of course there is a huge amount of work in the Matplotlib backend, and implementing one based on the object based Matplotlib API is not an immediate fix, but a wish for the long term. But I guess it's useful to keep that in mind.

In #27009 I had all sort of weird errors related to this. One of them was that some tests broke just by adding a from pandas import Series somewhere. I didn't find a proper fix, but just kept reverting things and looking for alternative ways until errors disappeared.

May be things will be better after we merge #28531. Not sure if in the case of Matplolib it'll make a difference, since there are several files of tests. But not parallelizing tests in the same file can help, since I guess tests in the same file are likely to affect the same parts of pyplot.

Hope it helps. If I misunderstood and you think talking to someone from Matplotlib or Azure may help, let me know.

@jbrockmendel
Copy link
Member Author

Not sure if I understand. Do you mean you can't reproduce them locally?

I can reproduce it locally, I just can't figure out how to fix it.

@datapythonista
Copy link
Member

I can reproduce it locally, I just can't figure out how to fix it.

I see. Had a look, but can't really tell. Not sure if you can open a separate PR for each change, and at least we'll know which one if causing the error.

@jbrockmendel
Copy link
Member Author

Separate PRs it is.

@jbrockmendel
Copy link
Member Author

Closing while I troubleshoot this locally.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants