Finer grained control over constrained layout

Community
1

I’m trying to automate the process of removing all space between subplots that share axes, and I’m struggling to achieve more fine-grained control over the constrained layout calculation. Would appreciate feedback if I’m missing something or confirmation that my observations are bugs or reasonable feature requests.

I can’t seem to, e.g., exclude tick labels or axes titles via set_in_layout, e.g., with

%matplotlib inline

import matplotlib.pyplot as plt
rc = {
    "xtick.direction": "in", 
    "ytick.direction": "in",
    "figure.constrained_layout.use": True,
    "figure.constrained_layout.h_pad": 0,
    "figure.constrained_layout.w_pad": 0,
    "figure.constrained_layout.wspace": 0,
    "figure.constrained_layout.hspace": 0,
    "axes.labelpad": 0,
}
with plt.style.context(rc):
    fig, axes = plt.subplots(2, 2, figsize=(6, 4), sharey=True, sharex=True)
    
    axes[0, 0].axis("off")
    axes[0, 0].set_in_layout(False)
    for artist in axes[0, 0].get_children():
        artist.set_in_layout(False)
    title = axes[1, 0].set_title("title")
    title.set_in_layout(False)
    
    for ax in axes.flat[1:]:
        ax.set_ylim(-0.1, 1.1)
        ax.set_xlim(-0.1, 1.1)
    fig.canvas.draw()

image
image1239×839 8.76 KB

(See this issue for the reason for the fig.canvas.draw call.)

Maybe I don’t understand what to expect of set_in_layout, because my attempt above to remove all elements in the upper-left panel from the layout has no effect. (Perhaps axes that are turned off should automatically be excluded from the layout calculation - that’s what I would expect, at least.)

I’d also love an automated way to prune ticks whose labels extend beyond the axes limits - is there one, or might this be a good feature proposal?