Hello!
I’m doing some source reconstruction, so eventually I plot the SourceEstimate object I get, in the usual way. By default, for this kind of plots, MNE uses pyvistaqt as the default backend. However, when I do that, I get an error:
self.color_list.remove("#7f7f7f")
ValueError: list.remove(x): x not in list
After a bit of research, this error seems to make sense. As described here, something inside matplotlib was recently changed to use colours defined with RGB tuples rather than with hex strings. And MNE didn’t get the memo, and still relies on the fact that it expects to find the string “#7f7f7f” in that spot.
I’m not familiar with the insides of MNE, but it appears that in my case the thing that will not work is only needed when using the pyvistaqt backend. So, if I set it to matplotlib instead, I get a plot that looks reasonable.
But it also seems that the pyvistaqt backend is a lot more powerful (in particular, the plot I’m making right now is a plot of ERP in source space, so it’s nice to be able to see how it evolves over time rather than at a specific timepoint, and it looks like that only works with pyvista?).
Are we just supposed to use only the matplotlib backend until someone fixes this bug, or is there a workaround that would give me the nice plots I want?
Thanks!