DOC: zeros, empty, and ones now have consistent docstrings #10679
Conversation
|
Note that some say |
|
As any sequences work, woudn't it be better as sequence? I always use tuple myself for that purpose, but the latter would be more accurate. |
|
There are other places where we try to encourage users to use a specific convention even though a more general one is allowed. It makes it easier for us to switch to a deprecation warning in future. Perhaps someone else should voice an opinion on that |
|
Fair enough -- I've included this change, and docstrings should be coherent on that as well now! |
|
@eric-wieser Since the PR is already open, should I do something? I'm not very familiar with GitHub workflow :( I want to note that additional change is needed in the signature of |
|
@godaygo Hmm, |
|
@charris Do you mean improving the documentation, or changing the argument to something more transparent? |
|
@NelleV I can't decide :) But its not a big deal, so let's get this in, we can switch things around if, and when, we deprecate the current behavior of Thanks. |
This patch simply unifies the docstrings for np.empty, np.ones, and np.zeros.
There are other inconsistencies relating to docstrings, that I decided not to fix (in matlib.py, with the same functions for matrices).
closes #10611