Mathieu Scheltienne

Thanks for tackling my comments, I'll try to continue the review as soon as possible!

@colehank Hi! Apologies for the slow review pace, my bandwidth is limited, I pushed 4 commits which: - Clean-up a couple of docstrings/imports - Remove the `model_path` argument which is...

Thanks, I'll try to have a look tomorrow! > the classes for MEGNET are ['brain/other', 'eye movement', 'heart beat', 'eye blink'], which do not exactly match the classes used in...

No need to apologizes, that's fine and it does seem to be environment issues. The online CIs don't suffer from those, so they will tell us if `pytest` is passing...

@colehank @adam2392 Almost a year after it was open, I had some times today to push it over the finish line. The changes were mostly: - Documentation, adding the proper...

I haven't followed lately the addition to MNE, so some of the information below might be outdated. That said, you can use the default `dev` as a "for everything" group,...

It seems there is also an SSL issue with [https://apt.packages.shiftkey.dev/](https://apt.packages.shiftkey.dev/). Running ``` wget -qO - https://apt.packages.shiftkey.dev/gpg.key | gpg --dearmor | sudo tee /usr/share/keyrings/shiftkey-packages.gpg > /dev/null ``` Gives me `gpg: no...

Thanks for the PR! > However, I noticed that some other functions, for example [filter](https://github.com/mne-tools/mne-python/blob/72faa3caa2239c6a4d973fc4cfd4a66a57ebcf00/mne/filter.py#L2444), have a similar behavior as get_data before, where the last sample of the "bad" annotation...

I prefer (3), IMO this problem seems to be a bug within our private code/functions; and there is no reason to complicate the public API to tackle an internal bug.

> It seems that the issue is actually in the behavior of the parameter reject_by_annotation in the raw.get_data function. Basically when you opt to reject by annotation, the last sample...