API reference

This page covers the public surface exposed at the top level (import kinextract / from kinextract import ...), grouped by the submodule that defines each name. Internal helpers (leading-underscore names) are not listed here but remain documented and importable from their defining submodule for advanced use.

Configuration

`kinextract.FitConfig`	All tunable parameters controlling a single LOSVD spectral fit.
`kinextract.load_config_from_toml`	Build a `FitConfig` from a TOML configuration file.

Top-level fitting API

`kinextract.run_spectral_fit`	Run the full non-parametric LOSVD spectral fit end-to-end.
`kinextract.fit_state_map_with_optional_clean`	Run the MAP fit, optionally with sigma-clipping.
`kinextract.build_parameter_xscale`	Build a per-parameter scale vector to precondition the L-BFGS-B fit.
`kinextract.run_iterative_clean_map`	Iteratively sigma-clip outlier pixels while refitting the MAP model.
`kinextract.compute_losvd_roughness`	Peak-normalised maximum absolute second difference of a LOSVD array.
`kinextract.compute_losvd_n_peaks`	Count topographically-prominent peaks in a LOSVD.

State

`kinextract.FitState`	Mutable container holding the full numerical state of one spectral fit.
`kinextract.precompute_losvd_interp`	Precompute linear-interpolation indices/weights from xl onto vgrid0.
`kinextract.precompute_ip_map`	Precompute the (npix, nlosvd) pixel-shift index map for convolution.
`kinextract.getnlosvd_fast_from_b`	Resample a trial LOSVD histogram onto the fine internal velocity grid.

Spectrum loading

`kinextract.load_spectrum_for_fit`	Load a galaxy spectrum file and return arrays trimmed to the fit window.
`kinextract.make_fit_state`	Load a spectrum and assemble the fully-populated FitState for one fit.
`kinextract.build_initial_guess_nonparam`	Build the initial non-parametric-LOSVD parameter vector and its bounds.
`kinextract.choose_norm_wavelength_frame`	Determine whether a `.norm` file's wavelength column is rest- or observed-frame, and return it converted to the frame the fit expects.
`kinextract.fortran_rebin_after_redshift`	Resample a de-redshifted spectrum onto a uniform wavelength grid.

Continuum

`kinextract.asymmetric_least_squares_continuum`	Fit a smooth asymmetric-least-squares baseline to a spectrum (Eilers 2003).
`kinextract.grow_boolean_mask`	Dilate a boolean mask by a fixed number of pixels on each side.
`kinextract.grow_boolean_mask_A`	Dilate a boolean mask by approximately a fixed wavelength on each side.
`kinextract.robust_sigma`	Robust standard-deviation estimate via the normalized MAD.

asymmetric_least_squares_continuum/robust_sigma are standalone, one-time pre-normalization utilities – not part of the main fitting pipeline.

Joint continuum-in-the-model fit

The continuum-cofitting method (FitConfig.fit_continuum = True). See kinextract.joint for the full rationale.

`kinextract.run_joint_fit`	Run the joint fit and package results in the same shape as `kinextract.fitting.run_spectral_fit()`'s `"outputs"` dict.
`kinextract.fit_joint`	Run the joint LOSVD + template + continuum-P-spline MAP fit.
`kinextract.fit_joint_auto_xlam`	Automatically select the joint model's LOSVD regularization strength.
`kinextract.fit_joint_auto_xlam_sigl0`	Self-consistent fixed-point refinement of the wing-taper's `sigl0`.
`kinextract.build_pspline_design`	Build a B-spline design matrix for the continuum.
`kinextract.evaluate_model_gp_joint`	Evaluate the joint forward model for a trial parameter vector.

LOSVD analysis

`kinextract.fit_losvd_gauss_hermite`	Fit the classic 4-moment Gauss-Hermite model to a non-parametric LOSVD.
`kinextract.fit_losvd_gauss_hermite_higher`	Fit a higher-order Gauss-Hermite model (h3..h_max_order) to a LOSVD.
`kinextract.gauss_hermite_losvd_model`	Four-moment Gauss-Hermite LOSVD model, legacy Fortran normalization.
`kinextract.gauss_hermite_losvd_model_ho`	Higher-order Gauss-Hermite LOSVD model, van der Marel & Franx (1993) convention.
`kinextract.fh3_fortran_like`	Third-order Gauss-Hermite basis polynomial, legacy Fortran normalization.
`kinextract.fh4_fortran_like`	Fourth-order Gauss-Hermite basis polynomial, legacy Fortran normalization.
`kinextract.getfwhm_fortran_like`	Estimate the full-width-at-fraction-of-max and flux moments of a curve.

Numerics

`kinextract.evaluate_model_gp`	Evaluate the forward model spectrum for a trial parameter vector.
`kinextract.objective_map`	Scalar objective function minimised by the MAP LOSVD/template fit.
`kinextract.objective_components`	Break down the MAP objective into its individual terms.
`kinextract.compute_weighted_template_spectrum`	Compute the weighted mixture of stellar templates for given weights.

Templates

`kinextract.interp_template_tp_with_outside`	Linearly interpolate a single template spectrum onto the galaxy grid.
`kinextract.build_template_matrix_fortran`	Build the full (npix, ntemplates) stellar template matrix.
`kinextract.resolution_mismatch_sigma_A`	Compute the Gaussian kernel needed to remove a data/template LSF mismatch.
`kinextract.convolve_gaussian_pixels`	Convolve flux with a Gaussian kernel of a fixed width in pixels.

Masking

`kinextract.build_emission_line_mask`	Build a boolean mask of emission-line pixels to exclude from the fit.
`kinextract.build_clean_protect_mask`	Build the boolean "protect" mask exempting known real features from clipping.

Plotting

`kinextract.plot_fit`	Plot the observed spectrum against the best-fit model and residuals.
`kinextract.plot_losvd`	Plot the recovered non-parametric LOSVD with its Gauss-Hermite overlay.
`kinextract.plot_continuum`	Diagnostic plot for a continuum-cofit fit (`kinextract.joint`).

Error estimation

`kinextract.LOSVDErrorEstimator`	Unified interface for estimating uncertainties on a recovered LOSVD.
`kinextract.estimate_losvd_errors`	Convenience wrapper: run all requested error estimation methods and return the combined summary.

Recovery-bias validation

`kinextract.assess_recovery_bias`	Measure empirical LOSVD recovery bias on a grid of known truths.
`kinextract.correct_recovered_losvd`	Apply an empirical bias correction from an `assess_recovery_bias()` table.
`kinextract.build_matched_mock`	Synthetic spectrum matched to st's instrument/template/continuum/noise.
`kinextract.true_losvd_on_grid`	Ground-truth LOSVD histogram on xl, normalized so `sum(b) == 1`.

I/O

`kinextract.read_galaxy_index_flux_err`	Read a raw 3-column galaxy spectrum file.
`kinextract.read_norm_spectrum`	Read a pre-normalized `.norm` spectrum file.
`kinextract.read_template_xy`	Load a single stellar/SSP template spectrum file.
`kinextract.read_template_list`	Read a template-list file and resolve each entry to a full path.
`kinextract.read_galaxy_params`	Read a legacy galaxy.params file.
`kinextract.write_fitlov_outputs`	Write the final LOSVD fit results to legacy-format ASCII files.
`kinextract.write_losvd_errors_file`	Write LOSVD vector with error bars to a file for modeling pipeline.
`kinextract.write_gh_errors_file`	Write Gauss-Hermite moment errors to a file.

Optional Bayesian (NUTS) path

Not part of the top-level namespace (requires pip install kinextract[bayesian]); import directly from kinextract.bayesian.

`kinextract.bayesian.fit_state_bayesian`	Run the full Bayesian LOSVD/template/continuum fit, called from `kinextract.fitting.run_spectral_fit()`.
`kinextract.bayesian.build_numpyro_model`	Build the NumPyro model function for one LOSVD/template/continuum- offset posterior, at fixed regularization strength `st.xlam`.