mne-python saving epochs drop log plot

✉️ Mailing List Archive (read-only)
1

Hi all,
Anyone know how to go about saving the figure associated
with mne.viz.plot_drop_log, if possible?

The closest I've gotten is a blank image/file using:
plot_drop_log(epochs.drop_log, subject=subj)
matplotlib.pyplot.draw()
fig = matplotlib.pyplot.gcf()
fig.savefig(fname)

thanks in advance

Kambiz
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.nmr.mgh.harvard.edu/pipermail/mne_analysis/attachments/20140124/6fa594e5/attachment.html

2

hi Kam,

with the current version of mne-python you can do:

epochs.plot_drop_log()
import matplotlib.pyplot as plt
plt.savefig('fig.pdf')

HTH
Alex

3

Hi Alex,
I've tried this as well with same result i.e., blank file. Here is what I
am doing:

import matplotlib.pyplot as plt
_ = epochs.drop_bad_epochs()
plot_drop_log(epochs.drop_log, subject=subj)
plt.savefig(info.get('filename')[:-4] + '_droplog', dpi=100, format='pdf')

I can verify this works because I see the figure window and appropriate
drop plot. The documentation says plot_drop_log returns a float, so I am
not sure where in the ether the associated figure window ends up. I am
using mne-python version 0,8.

cheers,
Kam

4

hi Kam,

plt.savefig(info.get('filename')[:-4] + '_droplog.pdf', dpi=100, format='pdf')

works fine on my machine.

Note that I added the .pdf extension.

Alex

5

Kam,

I've recently spotted a similar problem when trying to save matplotlib
figures from a ipython qt-console that runs in inline mode.

Can you make sure you use ipython with a qt backend?

Best,
Denis

6

Thanks Denis. I finally got this thing to work. For anyone who is
interested; I edited my matplotlib config file setting for backend to
use Qt4agg and set the bindings for the backend to pyside i.e., the
specific modifications I made in the matplotlibrc file are:

`backend: Qt4agg`
`backend.qt4 : PySide`

cheers,
Kam

P.s. The alternative to PySide is PyQt4, which does not appear to be on
good terms with the Anaconda python environment. I believe it is not
available for install through their repositories, and I was also unable to
find it via `pip`. PyQT4 depends on a module called SIP that installs
rather painlessly after downloading source and building, however I
encountered lots of ugly errors trying the same approach with PyQT4...TBC

7

Hi Kam,

great to know and thanks for sharing this.
Another way to do it is exemplified in my .bash_profile file :

https://gist.github.com/dengemann/8674149#file-bashrc-L31

and the following line

Cheers,
Denis