Species

- Added support for Python 3.12, dropped support for Python 3.9
- Separated atmosphere and disk luminosity in `ModelBox` parameters
- Fixed issue with finding missing grid spectra
- Fixed issue with plotting filter profile width
- Added `pyproject.toml` and removed `setup.py`
- Removed most version restrictions from `requirements.txt`
- Added `font_size` parameter to `plot_spectrum`
- Fixed pathlib error when adding grids with `add_custom_model`
- Updated flux calibration tutorial
- Many small improvements and overall code maintenance

- Several small fixes in the retrieval framework (thanks wbalmer!)
- Added `flux_units` parameter to `get_mcmc_photometry`
- Support for fitting multiple `disk_teff` and `disk_radius` parameters in `FitModel`
- Support for fitting the combined photometry of unresolved binary stars with different parallaxes
- Multiple improvements for fitting a combination of two spectra in `FitModel`
- Support for flux ratio prior for filters when fitting a binary system
- Check for NaNs in the uncertainties when adding a spectrum with `add_object`
- Updated `plot_posterior` for showing posteriors from binary systems
- Returning the individual binary components with `get_mcmc_spectra`
- Added support for coolTLUSTY Y dwarf spectra
- Tutorial for comparing data with a model grid of spectra
- Support for fitting separate `ism_ext` parameters for a binary system
- Support for PHOENIX Husser et al. (2013) models
- Major refactoring of `FitModel`, improved extinction accuracy, faster interpolation based on Teff prior
- Added optional `database_file` parameter to `SpeciesInit`
- Added `fixed_param` parameter to `add_samples`
- Added `tag` parameter and several improvements in `get_residuals`
- Changed the values added with `inc_loglike` to ln(L)
- Included chi2 and n_dof in `ResidualBox`
- Fixed indentation problem in `CompareSpectra`
- Robust implementation for filenames by using the `pathlib` instead of the `os` module
- Using `pooch` instead of `urllib` for downloading files
- Fixed bug in `get_samples` with reading normal priors
- Fixed bug with storing the log-likelihood after running `run_multinest`
- Fixed Windows issue with the Linder et al. (2019) isochrones
- Check for NaN and inf when calculating the goodness-of-fit statistic in CompareSpectra
- Exclude SpeX spectra that can not be flux calibrated when adding with `add_spectra`
- Updated dependency versions
- Overall code maintenance, many minor improvements, and updated tutorials

Another fix in `setup.py`! Adding a missing JSON file to the `package_data`.

Fixing issue with `setup.py` to include the JSON files in the package on PyPI

- Fixed issue with `max_prob` parameter in `plot_posterior`
- Included `uniform_priors` and `normal_priors` to `SamplesBox`
- Added `show_priors` parameter to `plot_posterior`
- Fixed issue with wavelength precision in Exo-Rem grid
- Added `verbose` parameter to `list_filters` method of `ReadObject`
- Implemented the `units` parameter in `plot_spectrum`
- Added `convert_from` parameter to `convert_units` function
- Extended the BT-NextGen grid to 5 mm
- Allowing to fix the parallax when fitting with `FitModel`
- Including the blackbody radius (calculated from `disk_teff`) in the posterior with `inc_luminosity=True` in `plot_posterior`
- Added the `add_simple_object` method to `Database` for retrieving data from the SIMPLE Database
- Support for providing a spectrum array directly to the `spectrum` parameter of `add_object`
- Retrieving the `f_sed` parameter for each cloud species individually by setting `global_fsed=False`
- Added `spec_res` attribute to `ModelBox`
- Support for Sonora Elf Owl L-, T-, and Y-type grids
- Fixed issue with appending mode for opening the HDF5 database
- Fixed issue with `package_data` in `setup.py`

- Improved package structure by including additional subpackages and modules
- The implicit import of classes and functions is removed from the initialization of the package, instead these should be explicitly imported by the user
- Started with implementing improved output printing, added `print_section` function in `core_util`
- Support for BT-NextGen subsolar metallicity grid (`tag='bt-nextgen-subsolar'`)
- Changed default transmission threshold from 0.05 to 0.01 in `SyntheticPhotometry`
- GJ 758 B included in `companion_data.json` (thanks mperrin!)
- Tutorial included for plotting companion properties
- Fitting a flux scaling and offset in `FitModel`
- Updated `add_model_grid` such that only spectra files are unpacked within the selected `teff_range`
- Added `unpack_tar` parameter to `add_model`
- Update links in tutorials and documentation pages
- Added `verbose` parameter to `available_models` and `list_companions`
- Fixed issue with `float` type returned by likelihood function
- Fixed issue with `np.inf` in the likelihood that can not be handled by UltraNest
- Added `setup_retrieval` method to `AtmosphericRetrieval`
- Added gradient P-T parameterization from Zhang et al. 2023 (thanks wbalmer!)
- Support for nested sampling with Dynesty in `AtmosphericRetrieval` with `run_dynesty` method (thanks wbalmer!)
- Refactoring of the prior function in `FitModel` so it is shared by the sampling routines
- Added `normal_prior` parameter to `FitModel` and deprecated the `prior` parameter in `run_multinest` and `run_ultranest`
- Added `resume` and `kwargs_multinest` parameters to `run_multinest`
- Added `resume` and `kwargs_ultranest` parameters to `run_ultranest`
- Spectra are no longer removed when rerunning `add_object` with different spectrum labels
- Changed `scipy.interpolate.interp1d` to `numpy.interp` which produces the same output but `interp1d` is considered a legacy function
- Added `calc_phot` parameter to `contrast_to_mass`
- Retrieving `log_p_base` with physical clouds instead of using the condensation profile
- Added `param_fmt` parameter to `plot_spectrum`
- Fixed issue with appending to database during multiprocessing
- Added `verbose` parameter to `get_median_sample`, `get_probably_sample`, and `get_compare_sample`
- Store priors from `FitModel` in the database results
- Removed `interp_method` from configuration so linear interpolation is used
- Added `rad_vel` and `vsini` as optional retrieval parameters in `AtmosphericRetrieval`
- Added support for nested sampling with Dynesty in `FitModel` with `run_dynesty` method
- Renamed the `spec_res` parameter in `add_model` to `wavel_sampling`
- Deprecated the `smooth` parameter in `ReadModel` and `ReadPlanck` since the `spec_res` parameter is sufficient
- Renamed the `get_spec_res` method in `ReadModel` to `get_sampling`
- Moved the `convert_model_name` function to the `model_util` module
- Added option to use `wavel_range` parameter without setting `wavel_sampling` in order to not resample but store a restricted wavelength range
- Increased dependency versions in `requirements.txt`
- Updated documentation and tutorials
- Many small enhancements and several minor bug fixes