Spectral plots in Jupyter Notebook

[pythonic2020]

Hello, I'm having trouble getting a spectral plot to display in a Jupyter Notebook. I'm running SPy 0.22.1 in an up-to-date Miniconda environment with notebook 6.1.5 and Python 3.8.6. I have not changed my matplotlib backend, and don't know if I can enable the required WX backend. My backend is probably causing my problem - I'm searching on how to enable the WX backend now. My data is in ENVI format with band wavelengths in the header file. This code:

%matplotlib notebook
import spectral.io.envi as envi
image = envi.open('hyper.hdr', image='hyper')
from spectral.graphics.spypylab import imshow 
view = imshow(data=image[1300:1800, 1700:2200], bands=(71, 84, 97),
 origin='upper', stretch=(0.02, 0.98), source=image, 
 title='New data - Bands 71, 84, 97/RGB')

displays data perfectly, and I can pan and zoom using matplotlib buttons. However, when I double-click on a pixel, nothing happens. I think a spectral plot should appear, but it does not.

If I enter this code in a new notebook cell:

zoom = view.open_zoom()

and ctrl+lclick in the original window, the zoom window updates as it should, so interactive communication between cells appears to be working.

Any ideas on how to get a spectral plot to display would be greatly appreciated.

Thanks in advance...

[pythonic2020]

Never mind. After installing wxpython and pyopengl, and using this code:

import matplotlib as mpl
mpl.use('WXAgg')
import matplotlib.pyplot as plt
#%matplotlib notebook

I got the spectral plots (and view_cube) to work. Be sure to disable the matplotlib magic. Beautiful! :-)

[tboggs]

I had no idea interactive features could work in a jupyter notebook. Glad you figured it out.

[pythonic2020]

Hopefully my posts will help people get SPy going in notebooks in the future. The %matplotlib widget magic works very nicely for zooming and panning in notebooks, and for quickly saving plots to file!

SPy has suddenly become very important for my work, and I am just learning to use it on hyperspectral image data. However, I am having trouble with some very simple yet vital operations that aren't fully explained with examples in the documentation. Would it be possible for you to answer a few quick questions, perhaps via email or some other communications method? Posting questions here probably isn't appropriate. Your help would be greatly appreciated!

[tboggs]

If they're questions that could be useful to others, then it's fine to ask them in issues on GitHub.

[pythonic2020]

Excellent, will do. Thank you in advance. I'm amazed at the capability of SPy, especially the documentation of referenced articles for algorithms.