Incompatibility with plt.subplots(..., sharey=True)

[sholderbach]

As the drawcall updates the ylim using the sharey=True setting in matplotlib.pyplot.subplots() is not really useful to compare the weights of different matrices.

MRE:

mtx = logomaker.get_example_matrix('logomaker_logo_matrix')
fig, axs = plt.subplots(2, sharey=True)
logomaker.Logo(mtx * 2, ax=axs[0]) # Require a larger dynamic range
logomaker.Logo(mtx, ax=axs[1]) # ylims are updated to the last displayed range

Output:

Expected behavior:

If multiple logos with a shared y or x axes are drawn, the range accomodates the largest drawing to allow comparison of the magnitude.

---

matplotlib version used: 3.5.2

[liliblu]

I'm also running into this problem.
I'm getting around it by setting sharey=True, then saving the lowest ymin and highest ymax via ax.get_ylim() for each plot and finally setting ax.set_ylim(ymin, ymax) at the end of creating the figure. Hope that helps you too.

[atareen]

Thank you @liliblu for the solution!

@sholderbach can you try the above please? If the solution works for you, please close the issue.

[jbkinney]

Thank you all for raising this issue, and sorry for the delay in addressing it. I've just fixed the auto-scaling in Logo() to account for the sister ax objects in each fig object that have coordinated limits via sharey=True or sharex=True (see lines 1058-1063 and 1072-1077 in Logo.py). I'm closing this issue now; please raise again if you find any problems.

[sholderbach]

Awesome, thanks!