pqc.config

Define configuration objects for the PQC pipeline.

This module centralizes tuning knobs used by the QC pipeline. Defaults are conservative and intended to be safe for a wide range of datasets. Prefer constructing explicit config objects rather than scattering literals across modules.

All configuration classes are frozen dataclasses, making them hashable and safe to share across runs.

See also

pqc.pipeline.run_pipeline: Consumes these config objects.

Classes

BadMeasConfig([tau_corr_days, fdr_q, ...])	Configure bad-measurement detection.
EclipseConfig([enabled, center_phase, ...])	Configure orbital-phase eclipse event detection.
ExpDipConfig([tau_rec_days, window_mult, ...])	Configure exponential dip detection.
FeatureConfig([add_orbital_phase, ...])	Configure feature-column extraction.
GaussianBumpConfig([enabled, ...])	Configure global Gaussian-bump event detection.
GlitchConfig([enabled, min_points, ...])	Configure global glitch event detection.
MergeConfig([tol_days, freq_tol_mhz])	Configure time/metadata matching.
OrbitalPhaseCutConfig([enabled, ...])	Configure orbital-phase based flagging.
OutlierGateConfig([enabled, sigma_thresh, ...])	Configure hard sigma gating for outlier membership.
PreprocConfig([detrend_features, ...])	Configure covariate-conditioned preprocessing for detectors.
RobustOutlierConfig([enabled, z_thresh, scope])	Configure robust (MAD-based) outlier detection.
SolarCutConfig([enabled, approach_max_deg, ...])	Configure solar-elongation event detection and handling.
StepConfig([enabled, min_points, ...])	Configure step-like offset detection.
StructureConfig([mode, detrend_features, ...])	Configure feature-domain structure tests and detrending.
TransientConfig([tau_rec_days, window_mult, ...])	Configure transient exponential-recovery detection.