Scale 1D curve to reference (compute-only)#

Summary#

Compute a normal robust-fit or uncertainty-weighted lognormal scale between two 1D curves.

Metadata#

Import path: modacor.modules.base_modules.find_scale_factor1d.FindScaleFactor1D
Source: src/modacor/modules/base_modules/find_scale_factor1d.py
Module ID: FindScaleFactor1D
Module version: 20261002.1
Keywords: scale, calibration, lognormal, 1D

Required data keys#

signal
Q

Modifies#

scale_factor: signal, uncertainties, units
scale_background: signal, uncertainties, units

Required arguments#

with_processing_keys

Default configuration#

{
  "background_output_key": "scale_background",
  "background_uncertainty_key": "propagate_to_all",
  "diagnostic_prefix": null,
  "fit_background": false,
  "fit_max_val": null,
  "fit_min_val": null,
  "fit_model": "normal",
  "fit_val_units": null,
  "independent_axis_key": "Q",
  "interpolation_kind": "linear",
  "require_overlap": true,
  "robust_fscale": 1.0,
  "robust_loss": "huber",
  "scale_output_key": "scale_factor",
  "scale_uncertainty_key": "propagate_to_all",
  "signal_key": "signal",
  "uncertainty_weight_key": null,
  "use_basedata_weights": true,
  "with_processing_keys": null
}

Argument specification#

Argument	Type	Required	Default	Dependency role	Description
`background_output_key`	str	No	scale_background	-	BaseData key to store the fitted background output.
`background_uncertainty_key`	str	No	propagate_to_all	-	Named uncertainty component used for the fitted background uncertainty.
`diagnostic_prefix`	str or NoneType	No	-	-	Optional prefix for scalar ‘_point_count’, ‘_x_min’, ‘_x_max’, and ‘_reduced_chi_square’ outputs.
`fit_background`	bool	No	False	-	Whether to fit a constant background offset.
`fit_max_val`	float or int or NoneType	No	-	-	Maximum x-value for the fit (in fit_val_units).
`fit_min_val`	float or int or NoneType	No	-	-	Minimum x-value for the fit (in fit_val_units).
`fit_model`	str	No	normal	-	Scale estimator: normal or lognormal.
`fit_val_units`	str or NoneType	No	-	-	Units for fit_min_val/fit_max_val if provided.
`independent_axis_key`	str	No	Q	-	BaseData key for the independent axis.
`interpolation_kind`	str	No	linear	-	Interpolation kind passed to scipy/numpy interpolation.
`require_overlap`	bool	No	True	-	Require overlapping x-range between reference and work data.
`robust_fscale`	float or int	No	1.0	-	Robust loss scale parameter.
`robust_loss`	str	No	huber	-	Robust loss function name for the fit.
`scale_output_key`	str	No	scale_factor	-	BaseData key to store the scale factor output.
`scale_uncertainty_key`	str	No	propagate_to_all	-	Named uncertainty component used for the fitted scale-factor uncertainty.
`signal_key`	str	No	signal	-	BaseData key for the dependent variable signal. Working and reference signals must have compatible units; fitting uses the reference units.
`uncertainty_weight_key`	str or NoneType	No	-	-	Named propagated uncertainty component used for weighting on both curves. Required for lognormal fitting; normal fitting combines components only when this is None.
`use_basedata_weights`	bool	No	True	-	Use BaseData weights when fitting.
`with_processing_keys`	list	Yes	-	-	Two processing keys: working curve then reference curve.

References#

DOI 10.1107/S1600577513030117

Notes#

The fitted scale uncertainty is stored under scale_uncertainty_key. Reduced chi-square is a goodness-of-fit diagnostic; it does not automatically inflate the formal fitted uncertainty.