bold attributes so we don't have to deal with trailing underscores #63
Conversation
…lso looks better.
|
That code is called from |
|
Bold formatting for attributes sounds fine to me, should look OK visually. |
|
I'm not sure what you mean by "methods get turned into links". I guess that depends on the template? Scikit-learn doesn't do that: Do you have a minimal example with methods that are links so I can test? I imagine that if a method contains a trailing underscore, that's also a problem. |
|
Probably it's about the method autosummary:: table immediately below the
heading, which contains the links.
|
|
No I meant actual Just tested this PR, looks like things still work so no issue. |
|
OK, in it goes. Thanks @amueller |
|
thanks. Sorry for being unresponsive. |
Eh? A couple of days doesn't count as unresponsive in my book. Or otherwise I should go make many handfuls of apologies right now:) |
update numpydoc to upstream, include fix for line numbers try to exclude numpydoc from unit test (why would we test those?!) don't run flake8 on the sphinxext folder. make attributes bold (and literal) so we don't have to worry about trailing underscores. numpy/numpydoc#63 actually exclude numpydoc from tests
update numpydoc to upstream, include fix for line numbers try to exclude numpydoc from unit test (why would we test those?!) don't run flake8 on the sphinxext folder. make attributes bold (and literal) so we don't have to worry about trailing underscores. numpy/numpydoc#63 actually exclude numpydoc from tests
Also looks better.
Fixes #62.
If the bold formatting is not for you, we can do backticks instead. I feel that attributes should be literals, not interpreted by sphinx.