Fix CI via catching a warning

[sappelhoff]

PR Description

fixes CI, see: https://github.com/mne-tools/mne-bids/actions/runs/3589485847/jobs/6041964853

I think this is because of mne-tools/mne-python#11344

Merge checklist

Maintainer, please confirm the following before merging.
If applicable:

• All comments are resolved
• This is not your own PR
• All CIs are happy
• PR title starts with [MRG]
• whats_new.rst is updated
• New contributors have been added to CITATION.cff
• PR description includes phrase "closes <#issue-number>"

[codecov]

Codecov Report

Merging #1106 (01c7f0d) into main (fa69665) will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff           @@
##             main    #1106   +/-   ##
=======================================
  Coverage   95.28%   95.28%           
=======================================
  Files          24       24           
  Lines        3900     3900           
=======================================
  Hits         3716     3716           
  Misses        184      184           

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

[larsoner]

Yes that's the PR that caused it, but I think MNE-BIDS should maybe internally squash these warnings using a record_warnings or so. It's weird to get a warning when using read_raw_bids that units have changed when read_raw_bids is doing the right thing (right?) by changing the channel types according to the channels.tsv or whatever.

For example, I don't like that we have to use verbose='error' here:

mne-tools/mne-python#11348

to fix

https://app.circleci.com/pipelines/github/mne-tools/mne-python/17161/workflows/bb0bdc62-4fa4-4b9c-a781-dc73ab855516/jobs/50797

It seems like mne-bids should just "do the right thing" with channel types and units, and the warning makes it seem to the end user like it potentially is not

[hoechenberger]

@larsoner alao see #1100, I believe that's exactly what you mean