This version fixes a bug with the print routine that misplaced the series legends on stacked charts.
It Also adds a new feature (finally got around to it Robin

A right click context menu item "Signal/noise" has been added for the main plot.
The S/N is estimated using two spectra of the object that are derived from two equal groups of exposures obtained in the same observing run.
For example if 6 exposures are available then one spectrum could use exposures: 1,3,5 and the other could use exposures: 2,4,6.
One spectrum is used to normalize the other and then a set number of points along the wavelength axis are used to estimate the noise and generate a plot of S/N versu wavelength.
Here is an example for VV Cep that has no smooth regions in the continum for noise evaluation.
First the two spectra derived from exposures 1,3,5 and 2,4,6.
And the generated S/N plot ( ignore the fact that y axis still says relative intensity)
The noise estimate used 100 points along the wavelength axis. Note that the noise estimate is very sensitive to any wavelength shifts of strong lines between the two sets of data.
The "noisy" nature of this S/N plot is due in part to the fact that VV Cep has a great many lines at this resolution.
Although it will require care to use and interpret I think this calculation will be useful.
Tim