Start replacing ACCEPTS table by parsing numpydoc. #11300
Conversation
|
Well I did need to change artist.py to update the regex matching... (I didn't need to update the docstring, though, indeed.) |
| Parameters | ||
| ---------- | ||
| labels : list of str | ||
| labels : List[str] |
There was a problem hiding this comment.
Did we somewhere agree to switch to type annotation syntax?
There was a problem hiding this comment.
Grepping for 'List[' shows that it's definitely already being used.
lib/matplotlib/contour.py
576: lw = self.get_label_width(self.labelLevelList[lmin],
577: self.labelFmt, self.labelFontSizeList[lmin])
591: self.add_label(xmin, ymin, rotation, self.labelLevelList[lmin],
592: self.labelCValueList[lmin])
1039: artists : List[`.Artist`]
1042: labels : List[str]
1810: hatches : List[str], optional
lib/matplotlib/collections.py
386: urls : List[str] or None
387: .. ACCEPTS: List[str] or None
lib/matplotlib/tight_layout.py
52: num1num2_list : List[int]
lib/matplotlib/widgets.py
515: labels : List[str]
518: actives : List[bool], optional
2484: button : List[Int], optional
lib/matplotlib/backend_bases.py
1257: callbacks : List[Tuple[callable, Tuple, Dict]]
2368: callbacks : List[Tuple[callable, Tuple, Dict]]
lib/matplotlib/axes/_base.py
1613: v : List[float] or one of the strings listed below.
There was a problem hiding this comment.
I'll accept this for now, even tough I would like to have it discussed and agreed upon. I'm not sure if this is really the best way. Whatever the policy is, it needs documentation. I'll leave the topic for #10225.
In most cases, we can just look for 'param : type' in the docstring. We let ACCEPTS still have the priority to handle weird cases such as set_xlim where we document the first arg as "bottom", but it's "bottom, top" for the purpose of the ACCEPTS table; resolving *that* problem is punted to later. Only removed a few ACCEPTS entries as a proof of principle.
|
Added test. |
| Parameters | ||
| ---------- | ||
| labels : list of str | ||
| labels : List[str] |
There was a problem hiding this comment.
I'll accept this for now, even tough I would like to have it discussed and agreed upon. I'm not sure if this is really the best way. Whatever the policy is, it needs documentation. I'll leave the topic for #10225.
| # Much faster than list(inspect.signature(func).parameters)[1], | ||
| # although barely relevant wrt. matplotlib's total import time. | ||
| param_name = func.__code__.co_varnames[1] | ||
| match = re.search("(?m)^ *{} : (.+)".format(param_name), docstring) |
There was a problem hiding this comment.
I'm not quite sure, what we want the exact sematics to be.
What about foo : int, optional, default: None?
Do we want just int or int, optional, default: None?
There was a problem hiding this comment.
These are all setters, so the first argument "should" never be optional or have a default. As always, there are exceptions (set_xlim(left=...)), but again the point is to handle these with the fallback ACCEPTS table.
I don't think it's worth adding more machinery to this.
| :class:`Artists` are of the same type) and it is your responsibility | ||
| to make sure this is so. | ||
| Initialize the artist inspector with an `Artist` or an iterable of | ||
| `Artist`\s. If an iterable is used, we assume it is a homogeneous |
There was a problem hiding this comment.
This make the tests fail for me: SyntaxError: invalid escape sequence \s Do I have an old version of something lying around?
In most cases, we can just look for 'param : type' in the docstring.
We let ACCEPTS still have the priority to handle weird cases such as
set_xlim where we document the first arg as "bottom", but it's "bottom,
top" for the purpose of the ACCEPTS table; resolving that problem is
punted to later.
Only removed a few ACCEPTS entries as a proof of principle.
PR Summary
PR Checklist