Disable clipping on axis title

[anirudhkrishnan2718]

Hello,

I want to set a common title for the first 3 out of 4 columns in each row.
The default settings with tight = True and hpad = 2 produce:

I can work around this by setting tight = False and hspace = 5 to get something reasonable:

Is there a way to accomplish this without having to turn off the ultraplot's tight layout?

Thanks
Anirudh

Code to produce plots:

f, axs = uplt.subplots(
    refwidth=1,
    nrows=3,
    ncols=4,
    xlabel="X Label",
    ylabel="Y Label",
    grid=False,
    sharex=1,
    sharey=1,
    abc=["A", "", "", "", "B", "", "", "", "C", "", "", ""],
    # hspace=5,
    # tight=False,
    margin=0,
)
for ax, c in zip(axs, 3 * (3 * ["C0"] + ["C3"])):
    ax.plot(np.random.random(100), color=c, ls="", marker="o", ms=1)

for idx, ax in enumerate(axs[:, 1], 1):
    ax.format(title=f"Here are 3 examples of scatter plots with blue dots in Row {idx}")

axs[:, 3].format(title="Red Dots")

[cvanelteren]

I think the issue is caused by the long title; if you set the title to something smaler it does not occur.

[anirudhkrishnan2718]

I want the long title to span the first 3 columns.
The best case scenario would be to have the ability to select a contiguous set of subplots and then assign a common title to them. Something like
(hypothetical) ax[:2, :2].format(common_title = 'Group of 4 example panels')
which would be independent of the title per axis:

for idx, ax in enumerate(ax[:2, :2], 1):
    ax.format(title = f"Individual title {idx}")

I understand if a common_title setting like this cannot be implemented because of the 'no nested GridSpecs' limitation.

[cvanelteren]

We don't have that feature yet see #651

[anirudhkrishnan2718]

Is the workaround in my original question the only way to achieve a title that spills past the subplot boundaries currently?

[cvanelteren]

You can create a dummy axis and then turn that off:

fig, ax = uplt.subplots(ncols=3, nrows=4, margin=0, refaspect=1)
ax[:, 0].format(abc=["A", "B", "C", "D"])
ax[:3].format(title="A")
overlay = fig.add_subplot(1, 3, (1, 2), frameon=0)
overlay.axis(False)
overlay.format(xticklabels=[], yticklabels=[])
for spine in overlay.spines.values():
    spine.set_visible(False)

[cvanelteren]

but I got a fix for you so this won't be necessary.