ENH: Improve report plotting and animate_topomap

[larsoner]

I don't like how long it takes to generate reports, so I was trying to speed it up a bit. In doing so, I realized animate_topomap wasn't working properly, and refactored it to reuse plot_evoked_topomap code. This makes it more feature-complete, and I fixed its behavior (with both blit=True and blit=False). I also made it so the report now has butterfly=True from the animate_topomap, which I think is nice. From this testing code:

MWE

import mne

sample_dir = mne.datasets.sample.data_path() / "MEG" / "sample"
evoked_path = sample_dir / "sample_audvis-ave.fif"
cov_path = sample_dir / "sample_audvis-cov.fif"

evokeds = mne.read_evokeds(evoked_path, baseline=(None, 0))
#fig, anim = evokeds[0].animate_topomap(butterfly=True, frame_rate=2)
report = mne.Report()
report.add_evokeds(
    evokeds=evokeds[:1],
    n_time_points=20,
)
report.save("report_evoked.html", overwrite=True, open_browser=True)

You can see it:

I expect some test breakages, will push fixes for those.

[larsoner]

blit=False is no longer needed for the example to render properly (compare to main, where it is needed, and doesn't look as good / consistent with evoked.plot_topomap).

Ready for review/merge from my end!

[larsoner]

@drammock this one should be good to go, too

[drammock]

seems weird to have a global report setting that doesn't apply globally. At a minimum this should be documented in the docstring for Report constructor, but I'd be curious to hear why it needs to be ignored now.

[larsoner]

I'll have to look... IIRC I did this because there are currently two ways of setting the size here (global val, and the val you pass to brain_kwargs) and to me it was surprising if I passed brain_kwargs(size=my_size) I would get an image that wasn't my_size. Like if we have img_max_res=(500, 500) and I pass brain_kwargs=dict(size=(800, 800)) should it generate the brain image at 800x800 then scale it down to img_max_res? I'll look into the actual behaviors here on main for this and other functions...

[drammock]

I can see why that would be surprising. But if someone knew there was a max image size setting for the whole report, and it wasn't applying to the brain figures, that also seems surprising.

[larsoner]

Okay changed the behavior to respect report.image_max_width, and documented it!