DOC: reduce warnings when building, and rephrase slightly #13002
Conversation
|
|
||
| .. automethod:: numpy.core._internal._ctypes.strides_as | ||
| :noindex: |
| @@ -5148,7 +5154,7 @@ | |||
| out : ndarray, None, or tuple of ndarray and None, optional | |||
| A location into which the result is stored. If not provided or `None`, | |||
| a freshly-allocated array is returned. For consistency with | |||
| :ref:`ufunc.__call__`, if given as a keyword, this may be wrapped in a | |||
| ``ufunc.__call__``, if given as a keyword, this may be wrapped in a | |||
There was a problem hiding this comment.
ufunc.__call__ has no docstring. Should it?
There was a problem hiding this comment.
I think __call__ is always an implementation detail, whatever it does must be described elsewhere (e.g. in the main docstring of the object).
| axis : int, optional | ||
| The axis in the result array along which the input arrays are stacked. | ||
|
|
There was a problem hiding this comment.
Adding these lines quieted a warning "Block quote ends without a blank line; unexpected unindent." but I don't really understand why
There was a problem hiding this comment.
yeah that doesn't make a lot of sense; probably a numpydoc or sphinx bug
|
@mattip How did you set the EDIT: Unresolved conversations perhaps? |
|
@charris it shows up as a choice when you open a PR, not sure since when that exists. Seems like a nice thing, although I cannot see any method of setting it later (and I guess only the person who opened it can unset it). |
8736683 to
4b4af1a
Compare
|
This brings doc building down to 14 warnings, caused by two problems
|
|
We are still using sphinx 1.7.9 since sphinx 1.8 breaks our template (search stops working) #12044 |
c23ec69 to
3c10a30
Compare
|
Why does only macOS fail with Edit: fixed in bad8274 |
|
In the broadcast autosummary, Edit: with that, I would like to get this in before someone commits a PR that will require rebasing |
numpy/doc/glossary.py
Outdated
| @@ -416,6 +433,12 @@ | |||
| Universal function. A fast element-wise array operation. Examples include | |||
| ``add``, ``sin`` and ``logical_or``. | |||
|
|
|||
| vectorized | |||
There was a problem hiding this comment.
nit: vectorization, then use :term:`vectorized<vectorization>` above. Fine to leave to a follow-up
There was a problem hiding this comment.
adopted, but also expanded the content a bit.
|
The last commit drifted into the area of content modification since I needed some headings for a reference. If it is too much I can remove it. |
03ea609 to
4e2dce3
Compare
4e2dce3 to
76099ad
Compare
|
Rebased, squashed most of the commits (except the last two about vectorization, which maybe go too much toward editing new text). |
|
CircleCI now makes devdocs with no WARNINGS (there are a few glitches still, but sphinx-build does not count them). |
|
Awesome! I think we need a change to keep things permanently warning-free right? https://github.com/scipy/scipy/blob/master/.circleci/config.yml#L41 |
|
Yes, I would prefer to do that in a separate PR in case it gets bogged down in the details |
|
LGTM too, merged. Thanks @mattip, and thanks @eric-wieser for reviewing! This is pretty awesome - for years we've oscillated between 100 and 800 warnings, now we can actually stay at zero. |
| to code in whichever paradigm they prefer. | ||
| to code in whichever paradigm they prefer. This flexibility has allowed the | ||
| NumPy array dialect and NumPy `ndarray` class to become the *de-facto* language | ||
| of mulit-dimensional data interchange used in Python. |
There was a problem hiding this comment.
Thanks, will fix in #13063
Reduces the number of warnings to around
50, still a WIP to see if I can get it down further.15.