In pintk, the "phase" axis label for pre-fit and post-fit appears on the left instead of on the right.

[abhisrkckl]

Both phase and time Y axes are shown for "pre-fit" and "post-fit". The phase axis is shown on the right, but its Y label appears on the left for some reason.

[dlakaplan]

which files did you use for this?

[abhisrkckl]

This is using one of the simulated datasets I generated for #1673. But the bug appears for any set of par/tim files.

[dlakaplan]

I don't see the same thing.

[abhisrkckl]

That's surprising. I see the error with the same data files.

My system information is as follows:

Created: 2023-12-19T09:34:32.624030
PINT_version: 0.9.6+697.g2d8e13d0
User: Abhimanyu Susobhanan (abhimanyu)
Host: abhimanyu-HP-Envy-x360-2-in-1-Laptop-15-fh0xxx
OS: Linux-6.2.0-37-generic-x86_64-with-glibc2.35
Python: 3.10.13 (main, Sep 11 2023, 13:44:35) [GCC 11.2.0]
endian: little
numpy_version: 1.26.0
numpy_longdouble_precision: float128
scipy_version: 1.11.3
astropy_version: 5.3.4
pyerfa_version: 2.0.1.1
jplephem_version: 2.20
matplotlib_version: 3.8.0
loguru_version: 0.7.2
Python_prefix: /home/abhimanyu/miniconda3/envs/pint-noise
PINT_file: /home/abhimanyu/Work/PINT/src/pint/init.py
Environment: conda
conda_prefix: /home/abhimanyu/miniconda3/envs/pint-noise

[dlakaplan]

This could be something to do with Tcl/Tk. Probably not worth a lot of digging from PINT since it seems to be hard to reproduce unless we can find other reports online.

[scottransom]

Note that I'm also seeing this bug on my Linux system. Thanks to @dlakaplan for pointing this out. I tried to fix it to no avail. The code looks correct. It is quite annoying. I was thinking about overwriting the yaxis label and simply explicitly adding text at the proper position in device coordinates, but I'm not sure how that would work on re-scaling the window etc. I might try it.

[dlakaplan]

I recently used https://matplotlib.org/stable/api/_as_gen/matplotlib.axes.Axes.secondary_yaxis.html#matplotlib.axes.Axes.secondary_yaxis instead of twinx - it was a bit easier. But I think it only works in the latest matplotlib and won't work with oldestdeps