Release of version 0.19.1 by jnothman · Pull Request #9607 · scikit-learn/scikit-learn

jnothman · 2017-08-23T00:11:26Z

For merging to 0.19.X upon release.

jnothman · 2017-08-30T05:29:56Z

Is it controversial to include check_memory in this otherwise-bug-fix release?

amueller · 2017-08-30T14:46:11Z

I'm fine with including it. It could be considered a bugfix.

lesteve · 2017-08-30T15:23:05Z

Is it controversial to include check_memory in this otherwise-bug-fix release?

Fine with me. You may want to add #9649 when/if it gets merged.

lesteve · 2017-09-04T08:01:44Z

Probably this would be good to add #9683, so that we ensure that 0.19.1 will pass the tests with the next numpy release.

lesteve · 2017-09-12T09:31:04Z

I think this would be great to add #9670 which makes the nose dependency optional.

jnothman · 2017-09-12T09:39:07Z

@lesteve for the benefit of contrib developers?

…

On 12 September 2017 at 19:31, Loïc Estève ***@***.***> wrote: I think this would be great to add #9670 <#9670> which makes the nose dependency optional. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#9607 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAEz63B3MpIjkKtjI4sDQ06TtPXbjbQiks5shk9agaJpZM4O_VYV> .

lesteve · 2017-09-12T09:43:05Z

@lesteve for the benefit of contrib developers?

Exactly, for example imbalanced-learn has moved to pytest for running the tests but currently still has a dependency on nose, which #9670 will remove.

jnothman · 2017-09-12T14:26:18Z

Done

amueller · 2017-09-12T16:09:34Z

sweet :)

amueller · 2017-09-12T16:19:36Z

Circle fails?

Exception occurred:
  File "/home/ubuntu/scikit-learn/doc/sphinxext/sphinx_gallery/gen_gallery.py", line 280, in sumarize_failing_examples
    "\n" + "-" * 79)
ValueError: Here is a summary of the problems encountered when running the examples

Unexpected failing examples:
/home/ubuntu/scikit-learn/examples/applications/plot_stock_market.py failed leaving traceback:
Traceback (most recent call last):
  File "/home/ubuntu/scikit-learn/doc/sphinxext/sphinx_gallery/gen_rst.py", line 472, in execute_code_block
    exec(code_block, example_globals)
  File "<string>", line 135, in <module>
  File "<string>", line 135, in <listcomp>
  File "<string>", line 24, in wrapper
  File "<string>", line 64, in quotes_historical_google
  File "/home/ubuntu/miniconda/envs/testenv/lib/python3.6/site-packages/numpy/lib/npyio.py", line 1775, in genfromtxt
    converters[i].update(conv, locked=True,
IndexError: list index out of range

amueller · 2017-09-12T16:19:46Z

Oh is that the google stock retry thing?

jnothman · 2017-09-12T21:55:53Z

It is, though the retry is included in 0.19.1....

lesteve · 2017-09-29T15:00:00Z

I pushed a fix for numpy dev (test only) and the whole Google finance commit saga, hoping that it will make Circle pass.

lesteve · 2017-09-29T15:16:52Z

And of course, now the conda update saga strikes again ...

amueller · 2017-09-29T18:14:07Z

the pleasures of CI....

lesteve · 2017-10-03T07:28:02Z

Thanks for adding #9623! About whats_new, I was thinking, given the reorganisation in separate files in master, that it is simpler to add the whats_new entries in this PR (rather than in each separate PR), and once 0.19.1 is released to update v0.19.rst in master.

jnothman · 2017-10-03T11:56:15Z

Yes :)

jnothman · 2017-10-09T09:25:49Z

Btw, @lesteve, one approach is to release this, then merge the entire branch into master, just to be sure nothing is missed from master.

codecov · 2017-10-18T23:10:31Z

Codecov Report

Merging #9607 into 0.19.X will increase coverage by <.01%.
The diff coverage is 89.05%.

@@            Coverage Diff             @@
##           0.19.X    #9607      +/-   ##
==========================================
+ Coverage   96.17%   96.17%   +<.01%     
==========================================
  Files         335      335              
  Lines       61873    62117     +244     
==========================================
+ Hits        59504    59744     +240     
- Misses       2369     2373       +4

Impacted Files	Coverage Δ
sklearn/decomposition/incremental_pca.py	`100% <ø> (ø)`	⬆️
sklearn/cluster/dbscan_.py	`100% <ø> (ø)`	⬆️
sklearn/decomposition/dict_learning.py	`93.23% <ø> (ø)`	⬆️
sklearn/cluster/birch.py	`99.56% <ø> (ø)`	⬆️
sklearn/decomposition/online_lda.py	`100% <ø> (ø)`	⬆️
sklearn/cluster/spectral.py	`94.05% <ø> (ø)`	⬆️
sklearn/decomposition/pca.py	`94.52% <ø> (ø)`	⬆️
sklearn/manifold/mds.py	`84.46% <ø> (ø)`	⬆️
sklearn/gaussian_process/gaussian_process.py	`74.4% <ø> (ø)`	⬆️
sklearn/manifold/spectral_embedding_.py	`84.96% <ø> (ø)`	⬆️
... and 53 more

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update ef5cb84...063eb15. Read the comment docs.

jnothman · 2017-10-19T00:50:42Z

Except for #9593 (cross_val_predict binary decision_function etc) and #9950 (docs for "good first issue" label), I feel this is ready. I have pre-emptively added what's new for #9593. A double-check would be welcome, @lesteve, @amueller, @ogrisel, @agramfort, @TomDLT...

TomDLT · 2017-10-19T08:39:36Z

#9932 is missing

amueller · 2017-10-19T21:06:01Z

merged #9593

amueller · 2017-10-19T21:06:48Z

(not cherry picked, though. Can you cherry pick, I'll try to check the PR).

jnothman · 2017-10-20T06:05:39Z

Ready to merge and upload when you are, @amueller.

qinhanmin2014

PEP8 error for your reference, also, Reiichiro Nakano (#9593) seems missing from contributors.

qinhanmin2014 · 2017-10-20T06:57:08Z

sklearn/model_selection/tests/test_validation.py

 from sklearn.utils.testing import assert_almost_equal
 from sklearn.utils.testing import assert_raises
 from sklearn.utils.testing import assert_raise_message
+from sklearn.utils.testing import assert_warns_message


duplicate import, remove either one

qinhanmin2014 · 2017-10-20T06:57:32Z

sklearn/model_selection/tests/test_validation.py

 from sklearn.utils.testing import assert_array_almost_equal
 from sklearn.utils.testing import assert_array_equal
 from sklearn.utils.testing import assert_warns
+from sklearn.utils.testing import assert_warns_message


duplicate import, remove either one

jnothman · 2017-10-20T07:15:22Z

Thank you!!

amueller · 2017-10-20T16:30:56Z

Ok I'll look and merge and release tomorrow. Deal?

amueller · 2017-10-20T20:05:23Z

How about #9635, #8742, #7373, #9897? not regression but potentially important bug fixes.

amueller · 2017-10-20T20:52:06Z

I pushed some more documentation fixes to your branch. let me know if you disagree with any of them. Apart from what I mentioned above, it looks pretty complete. I'll double check tomorrow that there's nothing in there that we don't want.

jnothman · 2017-10-21T10:02:53Z

The doc fixes are fine. I'm ambivalent about fixes to minor longstanding bugs like that to affinity propagation. including them in a minor release is more overhead than benefit.

jnothman · 2017-10-21T12:13:56Z

I've included #7373, #9897. Personally, I'd rather leave #9635, #8742 behind.

jnothman · 2017-10-21T12:14:01Z

merge at will.

amueller · 2017-10-21T22:33:49Z

ok cool! I think I agree with your assessment about the other bug fixes.

amueller · 2017-10-21T23:16:35Z

hm contributions are sorted by commits not alphabetically, though it looks like we did the same for 0.19. I kind preferred alphabetically but I don't mind enough to change it ;)

amueller · 2017-10-21T23:20:43Z

hm pulling ndcg and dcg is a breaking change between 0.19 and 0.19.1, right? That doesn't seem very nice? Shouldn't we deprecate and warn?

amueller · 2017-10-21T23:26:14Z

Lol have to merge manually to avoid squashing. I'm doing that now and tagging. I'd appreciate a summary of why removing these scores is ok, but I expect you and @agramfort thought this through.

amueller · 2017-10-21T23:59:18Z

builds are starting... working on conda-forge now
https://ci.appveyor.com/project/sklearn-wheels/scikit-learn-wheels
https://travis-ci.org/MacPython/scikit-learn-wheels

amueller · 2017-10-22T00:30:40Z

hm my internet is kinda dead. if anyone wants to do conda-forge, please go ahead.

amueller · 2017-10-22T01:06:11Z

Not sure if this is right conda-forge/scikit-learn-feedstock#60

amueller · 2017-10-22T01:06:35Z

also see ContinuumIO/anaconda-issues#6809

jnothman · 2017-10-22T03:09:18Z

thanks. ndcg was too broken for anyone to have successfully used it

jnothman · 2017-10-22T03:11:03Z

and I'd never noticed what's new was previously alphabetical!

amueller · 2017-10-23T03:18:38Z

When we removed the commit counts I wanted to do alphabetical. Need to push the wheels to pypi, rest looks done. Will to tomorrow morning.

jnothman · 2017-10-23T03:47:54Z

Oh, you mean that order. Ah. Right.

lesteve · 2017-11-08T11:00:43Z

Hmmm it looks like the nose hard dependency removal is not part of 0.19.1 😕 ... I am pretty sure I checked at one point and #9607 (comment) seems to say the same thing. It's weird also I am not able to see the diff or commit of this PR ...

jnothman · 2017-11-08T11:20:33Z

Sorry. I'm not sure how that happened. Well, I guess we'd better document it in what's new for 0.20.

lesteve · 2017-11-08T14:03:44Z

Oh well ... maybe something to keep in mind if there is a 0.19.2.

jakirkham · 2018-06-25T00:14:02Z

Should the associated milestone be closed?

lesteve · 2018-06-25T05:42:35Z

Thanks, I closed the 0.19.1 milestone.

jnothman force-pushed the for-0.19.1 branch from aa1dcb9 to 07e0a77 Compare September 12, 2017 14:27

jnothman force-pushed the for-0.19.1 branch from 07e0a77 to 7d6f307 Compare September 12, 2017 22:01

jnothman mentioned this pull request Sep 24, 2017

[MRG+1] TST Improve SelectFromModel tests #9733

Merged

jnothman mentioned this pull request Oct 1, 2017

[MRG+1] Changes default for return_train_score to False #9677

Merged

lesteve added this to the 0.19.1 milestone Oct 17, 2017

qinhanmin2014 reviewed Oct 20, 2017

View reviewed changes

amueller merged commit 0e1d261 into scikit-learn:0.19.X Oct 21, 2017

Uh oh!

Conversation

jnothman commented Aug 23, 2017

Uh oh!

jnothman commented Aug 30, 2017

Uh oh!

amueller commented Aug 30, 2017

Uh oh!

lesteve commented Aug 30, 2017

Uh oh!

lesteve commented Sep 4, 2017

Uh oh!

lesteve commented Sep 12, 2017

Uh oh!

jnothman commented Sep 12, 2017 via email

Uh oh!

lesteve commented Sep 12, 2017

Uh oh!

jnothman commented Sep 12, 2017

Uh oh!

amueller commented Sep 12, 2017

Uh oh!

amueller commented Sep 12, 2017

Uh oh!

amueller commented Sep 12, 2017

Uh oh!

jnothman commented Sep 12, 2017

Uh oh!

lesteve commented Sep 29, 2017 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

lesteve commented Sep 29, 2017

Uh oh!

amueller commented Sep 29, 2017

Uh oh!

lesteve commented Oct 3, 2017

Uh oh!

jnothman commented Oct 3, 2017 via email

Uh oh!

jnothman commented Oct 9, 2017

Uh oh!

codecov bot commented Oct 18, 2017

Codecov Report

Uh oh!

jnothman commented Oct 19, 2017

Uh oh!

TomDLT commented Oct 19, 2017

Uh oh!

amueller commented Oct 19, 2017

Uh oh!

amueller commented Oct 19, 2017

Uh oh!

jnothman commented Oct 20, 2017

Uh oh!

qinhanmin2014 left a comment

Choose a reason for hiding this comment

Uh oh!

qinhanmin2014 Oct 20, 2017

Choose a reason for hiding this comment

Uh oh!

qinhanmin2014 Oct 20, 2017

Choose a reason for hiding this comment

Uh oh!

jnothman commented Oct 20, 2017

Uh oh!

amueller commented Oct 20, 2017

Uh oh!

amueller commented Oct 20, 2017 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

amueller commented Oct 20, 2017

Uh oh!

jnothman commented Oct 21, 2017 via email

Uh oh!

jnothman commented Oct 21, 2017

Uh oh!

jnothman commented Oct 21, 2017

Uh oh!

amueller commented Oct 21, 2017

Uh oh!

lesteve commented Sep 29, 2017 •

edited

Loading

amueller commented Oct 20, 2017 •

edited

Loading

amueller commented Oct 21, 2017 •

edited

Loading