Merging FitPlanck in FitModel, support for calibration parameters, reading subsets of object data

[tomasstolker]

• Merged the FitPlanck functionalities into FitModel to reduce code duplication. One or multiple blackbody curves can be fitted by setting model='planck'.
• Deprecated FitPlanck without backward compatibility.
• Extended the support for fitting calibration parameters with FitModel, including an instrument-dependent flux scaling, error inflation, and model covariances through a correlation length and amplitude (see Wang et al. 2020). The likelihood function and already existing covariances are updated accordingly. See API documentation for details on how to include the calibration parameters and set the priors.
• Added the fit_corr parameter in FitModel which takes a list of spectrum names for which the covariances are fitted.
• Added type hints and type checks to some of the classes and methods.
• Changed the use of inc_phot and inc_spec in get_object, FitModel, and get_residuals. The changes are backward compatible so the argument can still be a boolean. However, it is also possible to provide a list with filter names to inc_phot or spectrum names to inc_spec if only part of the data should be read.
• Deprecated (without backward compatibility) the filters parameter in get_object, FitModel, and get_residuals. Selecting a subset of photometry is now done with inc_phot.
• Added a check in download_filter to ensure that only filter profiles from energy counting detectors are used.
• Added a smooth parameter in get_spectrum from ReadPlanck to smooth a spectrum with a Gaussian to a specified resolution. This is also used in FitModel.
• Several minor improvements, including new plotting functionalities.
• Updated the dependency versions were possible.