Public Member Functions
None	__init__ (self, StatisticConfig config, ObservableInterface observable_interface)

None	select_experiment (self, str experiment)

None	select_experiments (self, Sequence[str] experiments)

None	select_experiments_all (self)

bool	has_experiment_selection (self)

List[str]	selected_experiments (self)

None	select_experiment_observables (self, Sequence[ExperimentObs] observables)

None	select_experiment_observables_all (self)

bool	has_experiment_observable_selection (self)

List[ExperimentObs]	selected_experiment_observables (self)

None	reload_nuisance_specs (self)

None	set_nuisance_user_file (self, str\|Path user_yaml_path)

None	clear_nuisance_user_file (self)

None	update_cache (self, Optional[Sequence[ParamId]] p_specs=None)

Dict[BinnedObservableId, GaussianSummary]	compute_uncertainties (self)

MCResult	compute_uncertainties_and_sampling (self)

FitResultWithMaps	compute_MLE (self, Optional[Sequence[ParamId]] p_specs=None)

Contour	compute_confidence_contour (self, ParamId p1, ParamId p2, float z, Sequence[float] bounds, Optional[ContourOptions] options=None)

None	prepare_likelihood_for_scan (self, Optional[Sequence[ParamId]] p_specs=None)

None	set_manual_scan_point (self, Mapping[ParamId, float] p_hat, Mapping[ParamId, float] eta_hat)

LikelihoodScanGrid	scan_likelihood_around_current_point (self, ParamId p1, ParamId p2, float x_half_width, float y_half_width, int nx, int ny)

None	save_likelihood_scan_csv (self, str\|Path path, LikelihoodScanGrid grid)

Dict[ParamId, float]	get_all_obss_deps (self)

Dict[ParamId, float]	get_active_observable_dependencies (self)

Dict[ParamId, float]	get_p_specs (self, Optional[Sequence[ParamId]] p_specs=None)

Dict[ParamId, Dict[ParamId, float]]	get_all_correlations (self)

Dict[ExperimentObs, Dict[ExperimentObs, float]]	get_all_obs_correlations (self)

Dict[ExperimentObs, float]	get_obs_exp (self)

None	print_cache (self)

Public Attributes
	config

Protected Member Functions
	_to_cpp (self)

Protected Attributes
	_cpp

Detailed Description

High-level Python facade for statistical fits and scans.

Args:
    config: Statistic configuration. Python-only fields such as ``p_specs``
        and ``selected_experiments`` are used by this wrapper.
    observable_interface: Observable interface already configured with the
        observables to compute.

Examples:
    A typical workflow is::

        stat_cfg = StatisticConfig(MC_draws=1000)
        stat_cfg.p_specs = [my_param_id]
        stat = StatisticInterface(stat_cfg, observable_interface=obs_int)

        fit = stat.compute_MLE()
        print(fit.fit_ok, fit.p_hat)

        contour = stat.compute_confidence_contour(
            p1=my_param_id,
            p2=other_param_id,
            z=1.0,
            bounds=[-1.0, 1.0, -1.0, 1.0],
        )

Definition at line 438 of file StatisticInterface.py.

Constructor & Destructor Documentation

◆ init()

None StatisticInterface.StatisticInterface.__init__	(		self,
		StatisticConfig	config,
		ObservableInterface	observable_interface
	)

Create the C++ statistic interface and apply initial selections.

Definition at line 465 of file StatisticInterface.py.

Member Function Documentation

◆ _to_cpp()

StatisticInterface.StatisticInterface._to_cpp ( self )

protected

Return the underlying C++ statistic interface.

Definition at line 474 of file StatisticInterface.py.

◆ clear_nuisance_user_file()

None StatisticInterface.StatisticInterface.clear_nuisance_user_file ( self )

Return to the default user nuisance-specification path.

Definition at line 540 of file StatisticInterface.py.

◆ compute_confidence_contour()

Contour StatisticInterface.StatisticInterface.compute_confidence_contour	(		self,
		ParamId	p1,
		ParamId	p2,
		float	z,
		Sequence[float]	bounds,
		Optional[ContourOptions]	options = `None`
	)

Compute a two-dimensional confidence contour.

Args:
    p1: First fit parameter.
    p2: Second fit parameter.
    z: Gaussian-equivalent confidence level, for example ``1.0`` for an
        approximate one-sigma contour.
    bounds: Four values ``[xmin, xmax, ymin, ymax]`` defining the search
        rectangle.
    options: Optional contour/profiling options. Defaults to
        ``ContourOptions()``.

Returns:
    ``Contour`` wrapper exposing ``paths``, ``level`` and ``success``.

Raises:
    ValueError: If ``bounds`` does not contain exactly four values.
    Exception: Propagates backend failures, for example if MLE has not
        been computed before requesting the contour.

Definition at line 593 of file StatisticInterface.py.

◆ compute_MLE()

FitResultWithMaps StatisticInterface.StatisticInterface.compute_MLE	(		self,
		Optional[Sequence[ParamId]]	p_specs = `None`
	)

Compute the maximum-likelihood estimate for selected fit parameters.

Args:
    p_specs: Fit parameters. When omitted, ``self.config.p_specs`` is
        used.

Returns:
    Fit result containing best-fit values, profiled nuisances,
    uncertainties, correlations and minimum NLL.

Raises:
    Exception: Propagates C++ errors raised by cache construction,
        likelihood evaluation or minimization.

Definition at line 573 of file StatisticInterface.py.

◆ compute_uncertainties()

Dict[BinnedObservableId, GaussianSummary] StatisticInterface.StatisticInterface.compute_uncertainties ( self )

Propagate nuisance uncertainties and return Gaussian summaries.

Returns:
    Mapping from observable id to Gaussian or split-Gaussian summary.

Definition at line 554 of file StatisticInterface.py.

◆ compute_uncertainties_and_sampling()

MCResult StatisticInterface.StatisticInterface.compute_uncertainties_and_sampling ( self )

Run Monte-Carlo uncertainty propagation and keep raw samples.

Returns:
    ``MCResult`` containing raw samples, summaries and covariance.

Definition at line 565 of file StatisticInterface.py.

◆ get_active_observable_dependencies()

Dict[ParamId, float] StatisticInterface.StatisticInterface.get_active_observable_dependencies ( self )

Return dependencies currently visible to the Statistic manager.

This method is useful for runtime/lambda observables: it verifies that
dependencies declared on ``LambdaObservableConfig`` and propagated from
custom Wilson lambdas reached the statistic layer.

Definition at line 709 of file StatisticInterface.py.

◆ get_all_correlations()

Dict[ParamId, Dict[ParamId, float]] StatisticInterface.StatisticInterface.get_all_correlations ( self )

Return the nuisance correlation matrix as nested parameter maps.

Definition at line 727 of file StatisticInterface.py.

◆ get_all_obs_correlations()

Dict[ExperimentObs, Dict[ExperimentObs, float]] StatisticInterface.StatisticInterface.get_all_obs_correlations ( self )

Return experimental-observable correlations as nested maps.

Definition at line 731 of file StatisticInterface.py.

◆ get_all_obss_deps()

Dict[ParamId, float] StatisticInterface.StatisticInterface.get_all_obss_deps ( self )

Return all selected observable dependencies after nuisance pruning.

Definition at line 705 of file StatisticInterface.py.

◆ get_obs_exp()

Dict[ExperimentObs, float] StatisticInterface.StatisticInterface.get_obs_exp ( self )

Return experimental central values used by the statistic manager.

Definition at line 735 of file StatisticInterface.py.

◆ get_p_specs()

Dict[ParamId, float] StatisticInterface.StatisticInterface.get_p_specs	(		self,
		Optional[Sequence[ParamId]]	p_specs = `None`
	)

Return central values for selected fit parameters.

Args:
    p_specs: Fit parameters to query. Defaults to ``self.config.p_specs``.

Definition at line 718 of file StatisticInterface.py.

◆ has_experiment_observable_selection()

bool StatisticInterface.StatisticInterface.has_experiment_observable_selection ( self )

Return whether an exact measurement filter is active.

Definition at line 519 of file StatisticInterface.py.

◆ has_experiment_selection()

bool StatisticInterface.StatisticInterface.has_experiment_selection ( self )

Return whether an experiment filter is currently active.

Definition at line 498 of file StatisticInterface.py.

◆ prepare_likelihood_for_scan()

None StatisticInterface.StatisticInterface.prepare_likelihood_for_scan	(		self,
		Optional[Sequence[ParamId]]	p_specs = `None`
	)

Prepare and cache a likelihood object for manual two-dimensional scans.

Args:
    p_specs: Fit parameters used by the likelihood. Defaults to
        ``self.config.p_specs``.

Definition at line 638 of file StatisticInterface.py.

◆ print_cache()

None StatisticInterface.StatisticInterface.print_cache ( self )

Print the current C++ statistic cache for debugging.

Definition at line 739 of file StatisticInterface.py.

◆ reload_nuisance_specs()

None StatisticInterface.StatisticInterface.reload_nuisance_specs ( self )

Reload default and user nuisance specifications in the C++ manager.

Definition at line 527 of file StatisticInterface.py.

◆ save_likelihood_scan_csv()

None StatisticInterface.StatisticInterface.save_likelihood_scan_csv	(		self,
		str \| Path	path,
		LikelihoodScanGrid	grid
	)

Save a likelihood scan grid to a CSV file.

Args:
    path: Output CSV path.
    grid: Scan grid returned by :meth:`scan_likelihood_around_current_point`.

Definition at line 694 of file StatisticInterface.py.

◆ scan_likelihood_around_current_point()

LikelihoodScanGrid StatisticInterface.StatisticInterface.scan_likelihood_around_current_point	(		self,
		ParamId	p1,
		ParamId	p2,
		float	x_half_width,
		float	y_half_width,
		int	nx,
		int	ny
	)

Evaluate the likelihood on a rectangular grid around the current point.

Args:
    p1: First scanned parameter.
    p2: Second scanned parameter.
    x_half_width: Half-width of the scan range in the first parameter.
    y_half_width: Half-width of the scan range in the second parameter.
    nx: Number of grid points in the x direction.
    ny: Number of grid points in the y direction.

Returns:
    A ``LikelihoodScanGrid`` containing all evaluated points.

Definition at line 661 of file StatisticInterface.py.

◆ select_experiment()

None StatisticInterface.StatisticInterface.select_experiment	(		self,
		str	experiment
	)

Restrict the statistic manager to one experiment.

Args:
    experiment: Experiment name as stored in the experimental database.

Definition at line 478 of file StatisticInterface.py.

◆ select_experiment_observables()

None StatisticInterface.StatisticInterface.select_experiment_observables	(		self,
		Sequence[ExperimentObs]	observables
	)

Restrict statistics to exact experiment/observable/bin entries.

Args:
    observables: Exact measurements to retain. Each item combines an
        experiment label with a binned observable identifier.

Definition at line 506 of file StatisticInterface.py.

◆ select_experiment_observables_all()

None StatisticInterface.StatisticInterface.select_experiment_observables_all ( self )

Clear the exact experimental-observable selection.

Definition at line 515 of file StatisticInterface.py.

◆ select_experiments()

None StatisticInterface.StatisticInterface.select_experiments	(		self,
		Sequence[str]	experiments
	)

Restrict the statistic manager to a set of experiments.

Args:
    experiments: Experiment names to keep.

Definition at line 486 of file StatisticInterface.py.

◆ select_experiments_all()

None StatisticInterface.StatisticInterface.select_experiments_all ( self )

Clear any experiment selection and use all available experiments.

Definition at line 494 of file StatisticInterface.py.

◆ selected_experiment_observables()

List[ExperimentObs] StatisticInterface.StatisticInterface.selected_experiment_observables ( self )

Return the active exact measurement selection.

Definition at line 523 of file StatisticInterface.py.

◆ selected_experiments()

List[str] StatisticInterface.StatisticInterface.selected_experiments ( self )

Return the currently selected experiment names.

Definition at line 502 of file StatisticInterface.py.

◆ set_manual_scan_point()

None StatisticInterface.StatisticInterface.set_manual_scan_point	(		self,
		Mapping[ParamId, float]	p_hat,
		Mapping[ParamId, float]	eta_hat
	)

Override the scan reference point manually.

Args:
    p_hat: Fit-parameter values used as scan center/reference.
    eta_hat: Nuisance values used as scan reference.

Definition at line 648 of file StatisticInterface.py.

◆ set_nuisance_user_file()

None StatisticInterface.StatisticInterface.set_nuisance_user_file	(		self,
		str \| Path	user_yaml_path
	)

Use a custom user nuisance-specification file.

Args:
    user_yaml_path: Path to the YAML file overriding default nuisance
        specifications.

Definition at line 531 of file StatisticInterface.py.

◆ update_cache()

None StatisticInterface.StatisticInterface.update_cache	(		self,
		Optional[Sequence[ParamId]]	p_specs = `None`
	)

Refresh the C++ statistic cache.

Args:
    p_specs: Fit parameters to detach from the nuisance set. When
        omitted, ``self.config.p_specs`` is used.

Definition at line 544 of file StatisticInterface.py.

Member Data Documentation

◆ _cpp

StatisticInterface.StatisticInterface._cpp