Packaged V25 score-router artifact and runtime calibration helpers.

Package source: tabnetics.auto_router

Package overview

Packaged auto-router for tabnetics runtime calibration.

Stable exports

  • AUTO_ROUTER_ARTIFACT_VERSION (constant) - Source. Module-level constant exported by the package surface.
  • class AutoRouterOutput (class) - Source. Auto-router decision emitted by the packaged V25 score model.
  • class DescriptorOODGate (class) - Source. Robust per-dimension descriptor envelope persisted with router artifacts.
  • class ScoreExpandedRouter (class) - Source. Predict candidate BA/F1 scores and select a runnable tabnetics profile.
  • class ScoreRouterConfig (class) - Source.
  • def apply_router_output(config: Any, output: AutoRouterOutput) -> Any (function) - Source. Apply an auto-router decision onto a DFFSConfig-like object.
  • def compute_dataset_descriptor(X: Any, y: Any, metadata: Optional[Mapping[str, Any]] = None) -> Dict[str, Any] (function) - Source. Compute the dataset-only descriptor used by the packaged auto-router.
  • def default_artifact_path() -> Path (function) - Source. Return the bundled V25 router artifact directory.
  • def load_default_auto_router(model_dir: Optional[Path | str] = None) -> ScoreExpandedRouter (function) - Source. Construct a fresh router from the exact artifact bytes for each call.
  • def predict_auto_router(X: Any, y: Any, *, metadata: Optional[Mapping[str, Any]] = None, model_dir: Optional[Path | str] = None, descriptor_ood_gate_enabled: bool = False, crossfit_uncertainty_enabled: bool = False) -> AutoRouterOutput (function) - Source. Compute a dataset descriptor and emit the V25 auto-router decision.
  • CandidateUncertaintyPolicy (export) - Exported by tabnetics.auto_router.
  • CrossFitRouterUncertaintyArtifact (export) - Exported by tabnetics.auto_router.
  • RouterOutcomeRow (export) - Exported by tabnetics.auto_router.
  • RouterUncertaintyError (export) - Exported by tabnetics.auto_router.
  • fit_crossfit_router_uncertainty (export) - Exported by tabnetics.auto_router.
  • MISSINGNESS_DESCRIPTOR_KEYS (export) - Exported by tabnetics.auto_router.
  • MISSINGNESS_DESCRIPTOR_POLICY (export) - Exported by tabnetics.auto_router.
  • compute_missingness_descriptors (export) - Exported by tabnetics.auto_router.
  • missingness_descriptor_artifact_status (export) - Exported by tabnetics.auto_router.
  • missingness_descriptor_model_input_enabled (export) - Exported by tabnetics.auto_router.

Module details

tabnetics.auto_router.__init__

Source file

Packaged auto-router for tabnetics runtime calibration.

No top-level public symbols are exported directly from this module.

tabnetics.auto_router.runtime

Source file

Runtime auto-router for tabnetics pipeline configuration.

  • AUTO_ROUTER_ARTIFACT_VERSION (constant) - Source. Module-level constant exported by the package surface.
  • SCORE_ROUTER_ARTIFACT_TYPE (constant) - Source. Module-level constant exported by the package surface.
  • def compute_dataset_descriptor(X: Any, y: Any, metadata: Optional[Mapping[str, Any]] = None) -> Dict[str, Any] (function) - Source. Compute the dataset-only descriptor used by the packaged auto-router.
  • class DescriptorOODResult (class) - Source. Descriptor-space OOD evaluation for one auto-router decision.
  • class DescriptorOODGate (class) - Source. Robust per-dimension descriptor envelope persisted with router artifacts.
  • class AutoRouterOutput (class) - Source. Auto-router decision emitted by the packaged V25 score model.
  • class ScoreRouterConfig (class) - Source.
  • class ScoreExpandedRouter (class) - Source. Predict candidate BA/F1 scores and select a runnable tabnetics profile.
  • def default_artifact_path() -> Path (function) - Source. Return the bundled V25 router artifact directory.
  • def load_default_auto_router(model_dir: Optional[Path | str] = None) -> ScoreExpandedRouter (function) - Source. Construct a fresh router from the exact artifact bytes for each call.
  • def predict_auto_router(X: Any, y: Any, *, metadata: Optional[Mapping[str, Any]] = None, model_dir: Optional[Path | str] = None, descriptor_ood_gate_enabled: bool = False, crossfit_uncertainty_enabled: bool = False) -> AutoRouterOutput (function) - Source. Compute a dataset descriptor and emit the V25 auto-router decision.
  • def apply_router_output(config: Any, output: AutoRouterOutput) -> Any (function) - Source. Apply an auto-router decision onto a DFFSConfig-like object.

Documentation and webpages on this site are generated from authoritative internal sources using a combination of deterministic rules and generative AI. Errors are possible. Please report issues via GitHub Discussions or email [email protected].