MnePreprocessorRawToHaemo fixes and improvements

• fixed NIRS filter on hemoglobin concentration
• parametrized class for filters and partial pathlenght factor (ppf)

Wavelet Transform Coherence improvements

• added possibilities to set frequency_cuts instead of period_cuts for bins
• added choice between periods or frequency for y axis
• keep CWT for inspection
• plot using "jet" cmap by default
• fixed scientific display of frequencies on y axis

NIRS Recording improvements

• can now create a Recording from an already loaded mne.io.Raw object
• plot all the preprocessing steps with the CWT of a channel
• load label from nirs file instead of having to specify it manually
• added a mapping of sensors/detectors/channels to the closest Standard 1020 montage position

NIRS Dyad improvements

• set a new property is_intra on a dyad when both recording are the same object

XDF fixes

• properly load "nirs" type from NIRx Aurora LSL stream
• fixed issue when 2 markers stream were in the XDF file

Jupyter notebooks

• improved tutorial/fnirs_getting_started.ipynb
• deleted legacy tutorial/import_from_fnirs.ipynb
• new notebook tutorial/fnirs_recording_inspection.ipynb for in-depth inspection (quality control) of a recording being transformed to CWT, then intra-subject WTC