[ENH] Refactor dipole visualization to invert y-axis when min_freq > max_freq

[samadpls]

Added support for reversed y-axis in plot_tfr_morlet to allow min_freq > max_freq
Fixes #898

[samadpls]

@gtdang, should I update whats_new.rst for these changes?

[gtdang]

Yes, please do. Thanks!

[samadpls]

Thanks! I have updated the what's new file. @asoplata , I appreciate your review on the pull request.

[asoplata]

Apologies for the radio silence, @samadpls . We've been in emergency hotfix & feature development for several weeks due to needing updates to use in-class and elsewhere, but that should be over now. I have a personal TODO to review all of your PRs / etc. asap :). Inspecting this one now.

[asoplata]

Unfortunately there is an issue preventing this from working fully on the spectrogram display in the GUI (note that to see the spectrogram on the GUI for the default simulation at all, you have to apply #938 ). In the GUI, if you run the default simulation and generate a default spectrogram, everything works correctly. However, if I change the min and max frequencies such that min > max, no spectrogram is displayed. Simple changes to freqs here https://github.com/jonescompneurolab/hnn-core/blob/master/hnn_core/gui/_viz_manager.py#L304 aren't enough to fix the problem. Will try to investigate more tomorrow.