Fix signal scaling and add background fit to fitsignal plots

[luisfabib]

This PR fixes issues related to the plot() method of fitsignal's output. The experimental dipolar signals and the corresponding fit were being plotting with different scales seeming like one of the two was not being plotted.

Potentially closes #74

Changelog:

• Fitted dipolar signal and experimental data now are plotted with the same scales (the one of the original data).
• Uncertainty quantification of the dipolar signal fit now properly accounts for the scaling of the signal and displays properly.
• Background (scaled by its unmodulated contribution) is now plotted along with its uncertainty with the plot() method.
• Added option to request scaled background instead of the unscaled one as the fit.B output

[stestoll]

Instead of introducing a new input parameter to request the scaled background signal, it might be better to just return both unscaled and scaled background in the FitResult structure. Same for the overall signal. That is the easiest from the user perspective, as they don't have to remember yet another input parameter, and they can just look at FitResult and then pick what they want for plotting etc.

[luisfabib]

This would be an excellent idea, will do so. Indeed it is easier to search for outputs than to prepare inputs and options for certain behavior.

It would also go in the way of a recent request:

Couldn't we/you solve that in a way, where the output 'fits' to the input in a future version? Or was that fixed as well? :-)
Originally posted by @laenan8466 in #74 (comment)

[luisfabib]

The recent changes should now close #86 as well.