Add to `fit_peaks()` support for additional methods

[aphalo]

Currently the only method supported is spline, which is usually o.k. for estimating the wavelength location but not the height of a peak. Fitting Gauss, Lorentz and Voigt functions could be implemented as these are frequently used, at least when peaks are on a low baseline

See files in folder inst-not for background information. These would need to use nls() and some kind of self starting functions.

See also function fit_peaks() in package 'chromatographR'.

[aphalo]

Examples from Alan Morison

[aphalo]

This should be implemented at the same time #36 is addressed.

[aphalo]

See also (aphalo/ggpmisc#77). Currently stat_peaks() and stat_valleys() just find maxima and minima, either globally in the data set or locally within a moving window, and support the use of thresholds to avoid false positives (see aphalo/ggpmisc#70). However, once located, in many situations it is important to fit a function to data within the window to determine the possibly unobserved true location and value of maxima and minima and possibly other cardinal points describing the shape of peaks and valleys.

Package 'cardidates' provides an implementation for year-long time-series with time expressed in days. A more general approach would be useful in packages 'ggpmisc' and 'photobiology'.

Source code for 'cardidates' is available from a Subversion repository [svn checkout svn://svn.r-forge.r-project.org/svnroot/cardidates/](svn checkout svn://svn.r-forge.r-project.org/svnroot/cardidates/) and the released version is in CRAN.