Backport NumPy 1.13.0 fixes to 0.18.X #9137
Conversation
Until now we were in a edge case on assert_array_equal
numpy.apply_along_axis has changed behaviour when the function passed in returns a 2d array
|
Not sure why one Travis CI job failed. The others passed. There's no real indication in the failed job. Also it looks like some assertion in Maybe both of these could use a restart? |
|
There is a failing example in 0.18.1 which might make circle fail regardless
…On 16 Jun 2017 8:28 am, "jakirkham" ***@***.***> wrote:
Not sure why one Travis CI job failed. The others passed. There's no real
indication in the failed job.
Also it looks like some assertion in docutils went off in CircleCI.
Though I can't see why these changes would have caused that.
Maybe both of these could use a restart?
—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
<#9137 (comment)>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/AAEz6wWNqumWiu5Up2KfkCRXG-0Q4A_0ks5sEbASgaJpZM4N7xmQ>
.
|
|
I see. Could backport that patch as well. Which patch would that be? |
|
The change to plot_stock_market.py which used a deprecated Yahoo API
…On 16 Jun 2017 8:49 am, "jakirkham" ***@***.***> wrote:
I see. Could backport that patch as well. Which patch would that be?
—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub
<#9137 (comment)>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/AAEz6yYWHppWa-MIGdxk-aIzJMPgViS-ks5sEbTcgaJpZM4N7xmQ>
.
|
…n#9010) * DOC updated plot_stock_market.py to use Google Finance The implementations is intentionally very basic not to distract the users from the example. Specifically unlike ``quotes_historical_yahoo_ochl`` it does not cache downloaded data. I also had to remove some symbols because the have no data on Google for the specified date interval. These are WBA, LMT, KFT and MTU. Closes scikit-learn#8899 * DOC removed plot_stock_market.py from expected failing examples * Addressed review comments * Addressed another pass of review comments
|
Thanks. Guessing you are referring to this PR ( #9010 )? Have cherry-picked it into here. There was a conflict in |
|
There's another failure cropping up on CircleCI. ________________________________________________________________________________
../examples/cross_decomposition/plot_compare_cross_decomposition.py is not compiling:
Traceback (most recent call last):
File "/home/ubuntu/scikit-learn/doc/sphinxext/sphinx_gallery/gen_rst.py", line 467, in execute_script
exec(code_block, example_globals)
File "<string>", line 3, in <module>
NameError: name 'X_train' is not defined
________________________________________________________________________________ |
|
Maybe add 13cc121. There may be similar changes to examples in other commits ... Note that because sphinx-gallery is quite old in the 0.18.X you need to look only at the first error in examples and ignore the following ones. In particular for #9137 (comment), the first error is: ../examples/cross_decomposition/plot_compare_cross_decomposition.py is not compiling:
Traceback (most recent call last):
File "/home/ubuntu/scikit-learn/doc/sphinxext/sphinx_gallery/gen_rst.py", line 467, in execute_script
exec(code_block, example_globals)
File "<string>", line 11/home/ubuntu/scikit-learn/sklearn/decomposition/fastica_.py:116: UserWarning: FastICA did not converge. Consider increasing tolerance or the maximum number of iterations.
warnings.warn('FastICA did not converge. Consider increasing '
/home/ubuntu/scikit-learn/sklearn/decomposition/factor_analysis.py:224: ConvergenceWarning: FactorAnalysis did not converge. You might want to increase the number of iterations.
ConvergenceWarning)
, in <module>
TypeError: slice indices must be integers or None or have an __index__ method |
|
Failed on numpy master, possibly because |
|
Same issue on our master, I think we should fix. I'll ask the numpy folks what's up. |
|
Ah, the test failure on dev is here numpy/numpy#9185 but seems they want to fix it. |
|
Why is flake8 run on the non-python files? Is that fixed in master? hm.... that seems really strange? |
|
(I think the circle ci failure is due to the second) |
|
Merging these, then cherry-picking the others. |
|
Lol none of these has a whatsnew entry.... |
|
see #9167 for release branch |
|
Thanks for finishing this one out @amueller. |
Backports some PRs (linked below) needed for NumPy 1.13.0 support in
masterto0.18.X.xref: #7946
xref: #8040
xref: #8355
xref: #9010