Skip to content

MAINT: Fix Travis and Circle - #10540

Merged
ev-br merged 4 commits into
scipy:masterfrom
larsoner:cis
Jul 28, 2019
Merged

ev-br merged 4 commits into
scipy:masterfrom
larsoner:cis

Conversation

@larsoner

@larsoner larsoner commented Jul 28, 2019 •

Copy link
Copy Markdown
Member
  1. Bumps tolerance in a signal test so it should hopefully pass on OSX
  2. Fix link errors created by NumPy numpy.random namespace changes.

Closes #10535

@larsoner

Copy link
Copy Markdown
Member Author

Also pushed a commit to bump tolerances in special

@tylerjereddy tylerjereddy added the maintenance Items related to regular maintenance tasks label Jul 28, 2019
@tylerjereddy tylerjereddy added this to the 1.4.0 milestone Jul 28, 2019
@tylerjereddy

Copy link
Copy Markdown
Contributor

@larsoner Do you know if the doc issues in master mirror the ones I've observed when trying to backport in #10496? If so, let's label this for backporting to 1.3.x?

I guess I need to check the logs carefully.

A numpy.random.RandomState object for using the current state of the
created random generator container.
rand_state : `~numpy.random.mtrand.RandomState` object
A `~numpy.random.mtrand.RandomState` object for using the current state

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

What is ~ for?

Copy link
Copy Markdown
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Trims everything except what follows the last period in the ref (look at the rendered doc and you should see it)

@ev-br

ev-br commented Jul 28, 2019

Copy link
Copy Markdown
Member

LGTM since it makes the CI green again, modulo the explicit mtrand import.

I'd wait for a numpy dev to confirm they are fine with this before merging this though.

@rgommers

Copy link
Copy Markdown
Member

Yeah that doesn't look right. The objects are still there:

In [1]: np.__version__                                                              
Out[1]: '1.17.0'

In [2]: np.random.RandomState                                                       
Out[2]: numpy.random.mtrand.RandomState

In [3]: np.random.multinomial                                                       
Out[3]: <function RandomState.multinomial>

So I think it's a change that just affected intersphinx. CC @mattip FYI

@rgommers

Copy link
Copy Markdown
Member

IIRC it's still exposed in that namespace, just not documented in their API docs there anymore

ah that's why it doesn't link.

But do wonder if this is something numpy devs want to reconsider, e.g. add RandState back to their docs, with a corresponding note.

Indeed, I do think these should not have been removed from the docs, they're not even deprecated.

@rgommers

Copy link
Copy Markdown
Member

Well it's all still listed: https://numpy.org/devdocs/reference/random/legacy.html in the docs. It's too hard to find, and the main numpy.random docs reference mtrand explicitly which is wrong. So that should be reshuffled, and perhaps a :currentmodule: directive must be inserted in order to fix intersphinx.

For now though that won't unbreak our CI, that will take a 1.17.1 release I guess (that won't take very long, but more than a few days). So how about we merge this PR for now, fix the numpy issue, and open a scipy issue so we won't forget to revert this change before the next release?

@larsoner

Copy link
Copy Markdown
Member Author

Sounds good to me. Though I'm not sure we even need an issue, CircleCI will yell at us when the intersphinx inventory reverts with 1.17.1, if they do decide to change back.

@ev-br
ev-br merged commit 63481f9 into scipy:master Jul 28, 2019
@ev-br

ev-br commented Jul 28, 2019 •

Copy link
Copy Markdown
Member

OK, let's get this in, will open an issue in a moment. If CircleCI yells at us, great, we'll close it.

EDIT: Opened gh-10545

@ev-br

ev-br commented Jul 28, 2019 •

Copy link
Copy Markdown
Member

Thanks Eric, Ralf, all.

@tylerjereddy tylerjereddy added the backport-candidate This fix should be ported by a maintainer to previous SciPy versions. label Aug 4, 2019
@rgommers rgommers removed the backport-candidate This fix should be ported by a maintainer to previous SciPy versions. label Aug 7, 2019
@tylerjereddy tylerjereddy added the backport-candidate This fix should be ported by a maintainer to previous SciPy versions. label Jan 17, 2020
@tylerjereddy

Copy link
Copy Markdown
Contributor

Ah, I do need the tolerance fix for signal here on LTS branch.

tylerjereddy added a commit to tylerjereddy/scipy that referenced this pull request Jan 17, 2020
* update SciPy 1.2.3 LTS release notes to
account for a partial backport of scipygh-10540
@tylerjereddy tylerjereddy removed the backport-candidate This fix should be ported by a maintainer to previous SciPy versions. label Jul 31, 2020
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

maintenance Items related to regular maintenance tasks

Projects

None yet

Development

Successfully merging this pull request may close these issues.

TST: master branch CI failures

5 participants