ENH: make compute_covariance work with Evoked objects

[BabaSanfour]

Reference issue (if any)

Closes #14115.

What does this implement/fix?

Accept Evoked through existing epochs argument and reject keep_sample_mean=False. With Evoked data we treat samples will be treated as observations and channels as variables: spatial covariance over time, not an estimate of trial-to-trial noise covariance

Possible alternatives/Improvements

1. as an alternative we can wrap Evoked in Epoch object.
2. for arguments we accept inst (for instance instead of epochs) but this will require 2-3 releases and some backward compatibility maintenance.

Additional information

With the help of claude Opus5; all lines added were read and double checked.

[BabaSanfour]

@larsoner PR is ready; code works as intended, one improvement would be changing epochs to inst, this would require some deprecation cycles (?).

[wmvanvliet]

This is useful, also for the upcoming dipole fitting GUI :)

[BabaSanfour]

Thanks @wmvanvliet , i was going to make it an inst .. but not sure how i should handle deprecation here. Should i just change it ?

[wmvanvliet]

it's only not compatible if someone explicitly does mne.compute_covariance(epochs=epochs) which I don't think is many. We could add a named parameter epochs that will raise a deprecation warning when used but otherwise just maps to inst.

[wmvanvliet]

Thanks, @BabaSanfour!