Rampy

improvements

- we use the make_smoothing_spline function from Scipy for the GCVSmoothedNSpline smoothing mode (smooth function) and the gcvspline baseline, in replaced of the call to the gcvspline library. make_smoothing_spline is a reimplenetation of the Woltring Fortran code in gcvspline library, so modulo numerical errors, it returns the same results. This offers now the gcvspline baseline to all, without problem of Fortran compilation.
- bir are now an optional argument in rampy.baseline (because they do not appear in als, arPLS and drPLS algorithms)
- various improvements of the code formatting
- tests & examples updated

additions
- plot_spectrum(): makes an interactive plot of a signal, and also possibly of added baselines and smoothed signals.
- whittaker(): it is now possible to pass weights
- baseline(): a whittaker baseline going through regions of interest is also available in the baseline() function

- Correction in despiking function (thanks Kevin Yuan)

- fix installation : something went bad as we switched to pyproject.toml in 0.5.1

- fix installation dependencies

- map() function is now maps() to avoid any conflict with built-in Python map() function

- add the possiblity of importing lines of spectra for Horiba spectrometers

- add a despiking() function

- add external calibration in rameau

- add the posibility to calibrate K coefficient(s) on another dataset of spectra

- correction of a typo in mlclassificator function

- quick fix of a bug in read_horiba function (from rampy.maps)