MNE-Python channel unit warnings should be suppressed

[joe-rabbit]

The original issue was resolved, but a second issue emerged on this thread. Skip to this comment and read on from there. I've hidden the original problem in "details". (signed, @drammock)

Description of the problem

Hello, I am new to mne and have a problem with this dataset.

It shows
ValueError: invalid literal for int() with base 10: ''

Steps to reproduce

`datatype='eeg'
subject='024'
task='med1breath'
suffix='eeg'
bids_path=BIDSPath(subject=subject,task=task,suffix=suffix,datatype=datatype,root=bids_root)
sidecar_path=bids_path.copy().update(extension='.json')
raw = read_raw_bids(bids_path=bids_path)
`

Link to data

doi:10.18112/openneuro.ds003969.v1.0.0]

Expected results

Reads file

Actual results

Additional information

Platform: Windows-10-10.0.22000-SP0
Python: 3.9.7 (default, Sep 16 2021, 16:59:28) [MSC v.1916 64 bit (AMD64)]
Executable: C:\Users\josep\anaconda3\python.exe
CPU: AMD64 Family 23 Model 96 Stepping 1, AuthenticAMD: 16 cores
Memory: 15.4 GB

mne: 1.2.2
numpy: 1.20.3 {MKL 2021.4-Product with 8 threads}
scipy: 1.7.1
matplotlib: 3.4.3 {backend=module://matplotlib_inline.backend_inline}

sklearn: 0.24.2
numba: 0.54.1
nibabel: Not found
nilearn: Not found
dipy: Not found
openmeeg: Not found
cupy: Not found
pandas: 1.3.4
pyvista: Not found
pyvistaqt: Not found
ipyvtklink: Not found
vtk: Not found
qtpy: 1.10.0 {PyQt5=5.9.7}
ipympl: Not found
pyqtgraph: Not found
pooch: v1.6.0

mne_bids: 0.11.1
mne_nirs: Not found
mne_features: Not found
mne_qt_browser: Not found
mne_connectivity: Not found
mne_icalabel: Not found

[joe-rabbit]

What does it mean by this Warning.
and thank you it works!!

[mscheltienne]

I guess that the units have changed compared to older versions? I am not super familiar with this part, but I think it's linked to #1059
@hoechenberger I'll let you comment on this if you want ;)

[drammock]

when a channel type gets changed (in this case presumably from "EEG" to "Temp") then these warnings happen. Same with GSR1 and GSR2, which are skin conductance (in Siemens). All of them were apparently stored as "eeg" and MNE is changing their channel type based on the channel names, but it wants to make sure you know that in case it guessed wrong.

[joe-rabbit]

Oh got it thank you

[hoechenberger]

This issue should actually be transferred over to MNE-BIDS. I believe what's happening is in fact what @drammock said, except that the channel types aren't changed based on their name but based on the BIDS metadata. That is, I would think that we actually should suppress these warnings, as MNE-BIDS is doing exactly what it is supposed to do -- hence there is no need at all to emit a warning!

[welcome]

Hello! 👋 Thanks for opening your first issue here! ❤️ We will try to get back to you soon. 🚴🏽‍♂️

[drammock]

This issue should actually be transferred over to MNE-BIDS.

done

[sappelhoff]

Should probably be fixed here:

mne-bids/mne_bids/read.py

Line 621 in 0b1a61d

raw.set_channel_types(channel_type_bids_mne_map_available_channels)

• to take into account: Fix CI via catching a warning #1106

[hoechenberger]

It's now easy to suppress these warnings in MNE, and that change will be included in MNE 1.4.

Once MNE 1.5 (!) is out, we can safely switch over to that new functionality, as we intend to support the latest two versions of MNE in MNE-BIDS