Fitter

class src.binary_sed_fitting.Fitter

Collections of functions which deal with SED fitting.

  • Fitting blackbody spectrum

  • Functions to calculate chi2 values for a parameter grid.

  • Functions to find best fit values and their errors

  • At present, only chi2 minimization is implimented.

static blackbody(star, p0=[5000.0, -20])

Fit blackbody parameters to observational data using scipy.optimize.curve_fit.

Parameters:

  • star (Star object) – Star data with observed flux.

  • p0 (list, optional) – Initial guess for the fit parameters [Temperature, Log Scaling Factor].

Notes

Updates: Star.Te_blackbody, Star.log_sf_blackbody

static calculate_chi2(star, model, refit=True, plot=False, _trim=None, _minimising_param='chi2')

Calculate chi2 values for different model parameters.

Parameters:

  • star (Star object) – Star data.

  • model (Model object) – Model information.

  • refit (bool, optional) – Whether to refit the data (default is True).

  • plot (bool, optional) – Whether to plot the scatter matrix (default is False).

  • _trim (int, optional) – Number of top results to trim the DataFrame (default is None).

  • _minimising_param (str, optional) – Parameter over which to determine the best fit (‘chi2’, ‘vgf2’, ‘vgfb2’). Currently, only ‘chi2’ is implemented.

Notes

Updates: Star.df_chi2, Star.df_chi2_trimmed

  • Calculates chi2 values based on model parameters and observed data.

  • Saves the results into DataFrame star.df_chi2.

  • Optionally plots the scatter matrix using plot_chi2_matrix if plot=True.

  • Trims star.df_chi2 to top _trim results if _trim is provided.

  • Saves star.df_chi2 into a CSV file.

  • Uses calc_radius and calc_luminosity functions to calculate radius (R) and luminosity (L) based on model parameters.

Raises:

NotImplementedError – If _minimising_param is not ‘chi2’. (‘vgf2’ and ‘vgfb2’ are to be implemented).

static calculate_noisy_chi2(star, model, total_iterations=100, refit=True, plot=False, _percentile_threshold=10)

Calculate chi2 values for noise added data.

Parameters:

  • star (Star object) – Star data.

  • model (Model object) – Model information.

  • total_iterations (int, optional) – Number of iterations to perform noisy chi2 calculations (default is 100).

  • refit (bool, optional) – Whether to refit the data if a previous result exists (default is True).

  • plot (bool, optional) – Whether to plot the scatter matrix of noisy chi2 results (default is False).

  • _percentile_threshold (int, optional) – Percentile threshold for filtering chi2 values from the original chi2 results (default is 10).

Notes

Updates: Star.df_chi2_noisy

  • Uses star.df_chi2 to filter out top performing parameters based on _percentile_threshold.

  • Generates total_iterations versions of noisy observed data (da_obs_noisy) by adding Gaussian noise to the original flux.

  • Computes chi2 values for each noisy iteration and stores the best fitting parameters in star.df_chi2_noisy.

  • Saves star.df_chi2_noisy into a CSV file named based on star and model information.

Raises:

AttributeError – If star.df_chi2 does not exist or _percentile_threshold is not within the valid range [0, 100].

static get_blackbody_spectrum_loglog(log_wavelength, Te, log_scaling_factor)

Compute the log-log blackbody spectrum flux.

Adapted from the IUE RDAF 1989.

Parameters:

  • log_wavelength (float or list) – Logarithm of the wavelength in Angstroms.

  • Te (float) – Temperature of the source in Kelvin.

  • log_scaling_factor (float) – Logarithmic scaling factor to normalize flux (for a star: log10(radius/distance^2)).

Returns:

log_bbflux – Logarithm of the blackbody flux in ergs/cm2/s/A for the given temperature and wavelength range.

Return type:

float or list

static get_parameters_from_chi2_minimization(star, model)

Estimate best fit parameters from the least chi2 fit.

Parameters:

  • star (Star object) – Star data containing chi2 results and observed data.

  • model (Model object) – Model information used for fitting.

Notes

Updates: Star.data_all, Star.data, Star.data_not_fitted, Star.N_Np, Star.chi2_r, Star.vgf2, Star.vgfb2, Star.

This method retrieves and updates parameters from the best chi2 fit found in star.df_chi2.

It calculates additional statistical metrics for the fitted data.

It raises a warning if any data points have chi2 values significantly higher than the average (3-sigma threshold).

static get_parameters_from_noisy_chi2_minimization(star, model)

Estimate best fit parameters from noisy least chi2 fit.

Parameters:

  • star (Star object) – Star data containing noisy chi2 results.

  • model (Model object) – Model information used for fitting.

Notes

Updates: Star.data_all, Star.data, Star.data_not_fitted, Star.chi2_median, Star.chi2_r_median, Star.vgf2_median, Star.vgfb2_median, Star.<dim>_median, Star.<dim>_error_lower, Star.<dim>_error_upper

This method estimates the best fit parameters by calculating the median and errors from the noisy chi2 fits:

  • It uses Fitter.get_realistic_errors_from_iterations() to estimate realistic errors based on noisy fits and grid steps.

  • Updates various attributes in star related to best fit parameters, errors, and fit quality metrics.

  • Logs warnings if the best fit parameters differ from their median value.

static get_realistic_errors_from_iterations(parameter_values, grid_steps, component, para_type='parameter')

Estimates errors using spread in the noisy fits and edge cases.

  • If the best fit parameter is near a boundary:

    • Exaggerates errors and issues a warning.

  • If spread in iterations is less than step size:

    • Keeps errors similar to the step size.

  • Otherwise:

    • Calculates the 16nd, 50th, and 84th percentiles of the parameter_values.

    • Adjusts percentiles to grid_steps boundaries if needed.

    • Computes lower and upper bound errors in the parameter.

Parameters:

  • parameter_values (list) – List of parameter values obtained from noisy fits.

  • grid_steps (list) – List of grid steps available for the parameter.

  • component (str) – Component information related to the parameter.

  • para_type (str, optional) – Type of the parameter (default is ‘parameter’).

Returns:

  • para_50 (float) – Median of the parameter_values.

  • error_lower (float) – Lower bound error in the parameter.

  • error_upper (float) – Upper bound error in the parameter.

Notes

This method estimates realistic errors based on noisy fits and grid steps:

  • It calculates percentiles and adjusts them to grid_steps boundaries if necessary.

  • Issues warnings if the best fit parameter is close to model boundaries.

  • Handles cases where spread in iterations is minimal or zero.

static plot_chi2_matrix(df, model, alpha=0.3)

Plot scatter matrix for chi2 DataFrame.

Parameters:

  • df (DataFrame) – DataFrame containing chi2 values.

  • model (Model object) – Model information.

  • alpha (float, optional) – Transparency of points in the scatter matrix plot (default is 0.3).