Pandas-like plot shortcuts

[henryiii]

Pandas has several useful keyword arguments that wrap mpl functionality. We could make wrapper that provides this for all our functions. For example, for legend (also requested for at lest logy):

# For None, the check for get_legend_handles_labels would not throw an error, but just skip
def _plot_keywords(ax: Axis, legend: bool | None) -> None:
    "Anything added here must also be in pandas"

    if legend:
        if ax.get_legend_handles_labels()[0]:
             legend = ax.legend()
        else:
            raise <some error>

...
# For Stack, this defaults to legend=None
def plot(self, *, legend: bool | None = False, **kwargs: Any) -> Any:
        ...
        
       ret: Artist = ...
       ax = ret[0].ax
       hist.plot._plot_keywords(ax, legend=legend)

This is convenience only, this can be done by hand with a few lines. If we do add any of these keywords, this is how we should do it. Do we add this, though? Or keep to base mpl commands? The first step is making sure we have good examples of the base commands in our docs.

[LovelyBuggies]

@henryiii FYI, ax = ret[0].ax doesn't work (neither "axis" or "axes" works). Info about ret (len=3):

[StairsArtists(stairs=<matplotlib.patches.StepPatch object at 0x1842fbd90>, errorbar=<ErrorbarContainer object of 3 artists>, legend_artist=<ErrorbarContainer object of 3 artists>), StairsArtists(stairs=<matplotlib.patches.StepPatch object at 0x184342a60>, errorbar=<ErrorbarContainer object of 3 artists>, legend_artist=<ErrorbarContainer object of 3 artists>), StairsArtists(stairs=<matplotlib.patches.StepPatch object at 0x184363760>, errorbar=<ErrorbarContainer object of 3 artists>, legend_artist=<ErrorbarContainer object of 3 artists>)]

Wonder why there are 3 StairsArtists and how to obtain the axis.

[henryiii]

Looks like it's:

a[0][0].axes
# OR
a[0].stairs.axes
# OR
a[0].legend_artist[0].axes

The StairsArtist has several parts, including an error bar and a legend artist, so you need to get one of those (like the stairs itself) to get the axes.

[henryiii]

(I used dir(a[0]) to see what was allowed)