Codecov Report

Merging #8340 into master will increase coverage by <.01%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##           master    #8340      +/-   ##
==========================================
+ Coverage   94.74%   94.75%   +<.01%     
==========================================
  Files         342      342              
  Lines       60735    60801      +66     
==========================================
+ Hits        57543    57609      +66     
  Misses       3192     3192

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 c6b2002...9858e28. Read the comment docs.

can you give the link to the rendered docstrings on circle ci build?

http://scikit-learn.org/circle?8706/_changed.html

You missed one or two in RandomizedLogisticRegression and lasso_stability_path

I read a

default=np.finfo(np.float).eps

with the double back ticks on the html. I am not sure sure it's expected

cf https://8706-843222-gh.circle-artifacts.com/0/home/ubuntu/scikit-learn/doc/_build/html/stable/modules/generated/sklearn.linear_model.RandomizedLasso.html

http://scikit-learn.org/circle?8756/_changed.html

I don't get how this relates to #7793 which does not mention default values.

We sometimes indicate defaults, but rarely do so like this, which accords with numpy, etc.

Makes sense, better to close this pull then.

I don't mind adding some defaults in when it's useful. An example of when it is particularly not useful is when the default is something like None which is semantically unspecified. In this case , optional often tells you as much as you need to know, or otherwise adding (default) to the relevant part of the description (not the type spec which you are currently modifying).

Closing for now.

Noting here that the default for normalize is currently wrong in the master branch, if someone feels like fixing that.

Closing for now.

As you wish.