I recently got a warning raised by plot_compare_evokeds that I somehow couldn't switch off using mne.utils.use_log_level(0):

:50: RuntimeWarning: Only 1 channel in "picks"; cannot combine by method "mean".
cmap=cmap,
:50: RuntimeWarning: Cannot find channel coordinates in the supplied Evokeds. Not showing channel locations.
cmap=cmap,

That led me down a bit of a rabbit hole and I saw that within the code base, there are several ways how warnings are raised. Is each of those ways desirable for the particular situation? Or is that something to be unified, to consistently make use of a feature (such as setting a log level that all warning calls respect).

Examples:

• mne.utils.warn versus mne.utils.logger.warning in brainvision.py, misc.py, and kit.py
• some other modules use warnings.warn (fixes.py, _coreg_gui.py)

--> should these all be just mne.utils.warn instead?

PS: I also didn't solve that issue of not being able to suppress the warning yet, so if somebody has an idea, please let me know. Here is a MWE:

import numpy as np
import mne

# create data
sampling_freq = 200
times = np.linspace(0, 1, sampling_freq, endpoint=False)
sine = np.sin(20 * np.pi * times)

info = mne.create_info(ch_names=['sine'],
                       ch_types=['misc'],
                       sfreq=sampling_freq)

data = np.array([[0.2 * sine],
                 [0.4 * sine],
                 [0.6 * sine],
                 [0.8 * sine],
                 [1.0 * sine]])
evoked_array = mne.EvokedArray(data.mean(axis=0), info, tmin=-0.5,
                               nave=data.shape[0], comment='simulated')

# This will still raise warnings ...
with mne.utils.use_log_level(0):
    mne.viz.plot_compare_evokeds([evoked_array] * 2, picks=0, combine="mean")