feature request -- can brokenaxes be modified to act as a wrapper to existing plots?

[PhillipMaire]

Hey thanks for your work on brokenaxes! I was wondering if it is possible to make a wrapper for only basic operations that would work like this.

fig, ax = plt.subplots()
ax.plot(range(10), range(10))
baxes_wrapper = brokenaxes_wrapper(xlims=((0,2),(4,7)), ylims=((0,2),(4, 7)))
fig, ax = baxes_wrapper(fig, ax) # adds break to plot

the reason for this is that I plot using pandas dataframe and pivot tables and groupby functions which makes it hard to use brokenaxes as the plotter directly. thank you!

[bendichter]

@PhillipMaire thanks for the question. I did try to do something like that a few years ago, but it would require essentially copying all of the plot components to a new figure and axes, and I hit some roadblocks. In theory you should be able to do something like:

import pandas as pd
from brokenaxes import brokenaxes

df = pd.util.testing.makeMixedDataFrame()

bax = brokenaxes(xlims=((1,2), (3,4)))
for ax in bax.axs:
    df.plot(ax=ax)

but I just tried that and got a strange bug about exceeding the maximum ticks.

[PhillipMaire]

interesting it has no problem when passing index 1, only issues are with index 0 ???? 😕
Edited to make more sense

import pandas as pd
from brokenaxes import brokenaxes
df = pd.util.testing.makeMixedDataFrame()
bax = brokenaxes(xlims=((1,2), (3,4)))
df.plot(ax=bax.axs[1])

[PhillipMaire]

also y seems to work fine?!?! Just the first axis of the x axis do you think there may be something different in your code that you may have done to the first axis of the x?? this could potentially be all that is wrong

import pandas as pd
from brokenaxes import brokenaxes
df = pd.util.testing.makeMixedDataFrame()
bax = brokenaxes(ylims=((1,2), (3,4)))
df.plot(ax=bax.axs[0])
df.plot(ax=bax.axs[1])

[bendichter]

Good job with the investigation so far! Keep going and you'll find the issue.

[PhillipMaire]

I have looked through the code and I cant find any reason for this using the method above the y axis version is fully functional. I have compared the attributes of the working df.plot(ax=bax.axs[1]) and broken axes df.plot(ax=bax.axs[0]) and using a text compare I can't find anything different other than what is expected to be different (e.g. the actual numbers for the axes etc). I can only find this work around by making 3 spits and only plotting the 2nd and the 3rd one. there might be a way to remove the other parts of the first subplot but I'm not entirely motivated to do so 😪

import pandas as pd
from brokenaxes import brokenaxes
df = pd.util.testing.makeMixedDataFrame()
bax = brokenaxes(xlims=((0, 0),(0,2), (3,4)))

fig = plt.figure()
df.plot(ax=bax.axs[1])
df.plot(ax=bax.axs[2])
bax.axs[0].set_texts = []

[bendichter]

Did you try this approach with your own data? You may not run into this weird bug.