A major updates of S<sup>3</sup>Fit.
(1) the linear fitting in log space is revised;
(2) an annealing optimization is included to provide a rough global solution, which is transferred to non-linear least-square optimization for fine tuning;
(3) the estimation of calibration error is updated with an iterative process;
(4) the convolution with variable kernel width is supported to account for a wavelength-dependent spectral resolution;
(5) the output formats are revised to provide a more user-friendly access.
- The manuals for basic and advanced usages as well as the example Jupyter Notebook are updated to reflect the latest version.
- An additional manual page is added for a full list of available parameters of S<sup>3</sup>Fit.
- The flowchart of S<sup>3</sup>Fit's fitting pipeline is updated, while the strategy page is still being edited.
**Full Changelog**: https://github.com/xychcz/S3Fit/compare/v2.0.4...v2.2.0