matplotlib backend definition for live plotting

[ozumana]

Hi @SamSchott

How can I specify Qt5Agg as backend for Matplotlib (as required for live plotting)?

I am running one of your usage examples, which drives flawlessly my 2636B... the only problem is that plotting is not updating... and runs with the warning message below.

C:\Anaconda3\envs\MSS\lib\site-packages\keithley2600\result_table.py:766: UserWarning: Matplotlib is currently using module://ipykernel.pylab.backend_inline, which is a non-GUI backend, so cannot show the figure.

Thanx

[samschott]

In an IPython console, this is as simple as running %gui qt. You can also programatically choose your backend with matplotlib.use('Qt5Agg') before importing any plotting modules.

More information is available in the Matplotlib docs.

Hope this helps, Sam

[ozumana]

Thank you for the simple fix... however, I note that a plt.pause command was necessary in my code to ensure that the plot live refresh was visible.

[samschott]

Yes, you'll need to do that if your are adding data in the main thread. The GUI needs some time to do the actual plotting and plt.pause gives it that time.

I have added notes on this to the documentation.

[SirNate0]

On Windows I found I had to manually switch to that back end (on Linux I guess it was the default). An oddity I encountered in this was that I had to do any one of the following to get it to work:

1. Call matplotlib.use('Qt5Agg') before I imported matplotlib.pyplot. If I did that after the pyplot import, then there would be no warning message, but the plot would not update (even when saving the figure - just an empty graph saved).
2. Manually call the update() function on the ResultTablePlot
3. Use plt.switch_backend('Qt5Agg') instead of matplotlib.use (obviously after importing pyplot as plt).