Repository navigation
doc: add pandas and xarray fixtures to testing API docs - #28879
Conversation
|
Does not render for some reason. |
|
it's rendered here, but getting lost: Was wondering if I should maybe add sample usage to the docstrings |
|
That's what I meant. For some reason autofunction does not produce content here. If nothing else helps, you could resort to a manual |
|
Theory: I suspect the reason is that You may try autodoc_mock_imports but I don't know whether that can mock decorators (if not you will still fail when the import ties to run the |
ee76ec4 to
cf4e812
Compare
timhoffm
left a comment
There was a problem hiding this comment.
Mock import works 👍
Was wondering if I should maybe add sample usage to the docstrings
This would be very helpful.
|
@story645 you can add the usage examples or merge immediately - as you see fit. |
|
I'm locally building the usage examples, will push in a bit |
3534879 to
291541a
Compare
|
I so don't understand how it's even more broken now |
timhoffm
left a comment
There was a problem hiding this comment.
"Example" is not a numpydoc section. It's the plural https://numpydoc.readthedocs.io/en/latest/format.html#examples
Maybe this is tripping up numpydoc/autodoc?
|
how would you all feel about more pre-commits: https://numpydoc.readthedocs.io/en/latest/validation.html |
|
Well, it currently outputs a 27K-line table of issues on just the files in |
I know linting fix PRs aren't folks faves, but this seems like a good low hanging fruit task for sprints |
291541a to
16d15c7
Compare
16d15c7 to
aeb40be
Compare
Co-authored-by: Tim Hoffmann <2836374+timhoffm@users.noreply.github.com>
aeb40be to
dc3bdf5
Compare
Fine as long as the resulting PRs are trivial to review (small enough, only one kind of change at a time). I suspect the 27k-lines of issues contain a lot of aspects that we cannot or don't want to fulfill. I think you'd have to start with a |
…8879) Co-authored-by: Tim Hoffmann <2836374+timhoffm@users.noreply.github.com>

PR summary
Motivated by #28830, figured it'd be a lot easier for contributors to use/reviewers to suggest if the pandas and xarray fixtures were exposed in the docs. I'm not fond of the subheading title but also can't think of a better one.
PR checklist