Richard Höchenberger

I'm not sure this is the right approach (in terms of, where to do the conditional branching)… @larsoner Could you kindly take a look?

Thanks for your help here, @larsoner, this is much appreciated! I'd like to hear @SophieHerbst's input as well, esp. since she said they were struggling with disk space usage and...

With #1204, we'll only need to tag the version in the VCS for a release. Just saying that this needs to be reflected in the release protocol as well – ...

> We should also add an "automated workflow to publish to pypi" upon making a GH release based on a tag. Maybe it's a good time to tackle this now...

> @hoechenberger we are currently using mne_bids' interactive inspect_dataset() to remove noisy segments, but this is before epoching. I was thinking of doing the same, but in our experiments, we...

To recap, the problem is that you may have channels `'foo'` and `'bar`' in your raw data, but the channel order is `'bar'`, `'foo`' in channels.tsv When reading, MNE-BIDS tries...

@larsoner I edited my above comment to add more clarity as to how I understand the issue, could you please check if this description makes sense to you?

I think we have an ambiguous situation here in case we have an order mismatch between raw and channels.tsv In my mind we would keep the original order in raw...

> You could mean `raw.rename_channels({raw_ch_name: csv_ch_name for raw_ch_name, csv_ch_name in zip(raw.ch_names, csv['ch_names']})` which will do the wrong thing. Would it? I think it's one of two valid approaches. In fact,...

> Thinking about `raw.ch_names = ['MEG0111', ..., 'EEG001', ...]` but for whatever reason the person in their `channels.tsv` ordered EEG before MEG, the `rename_channels` approach would be wrong. Yes, but...