cleverly.Replayability#

class cleverly.Replayability(summarize_existing_artifacts: bool, retarget_cached_nuisances: bool, evaluate_stored_representer: bool, refit_nuisances: bool, evaluate_new_data: bool, unreconstructible: tuple[str, ...] = ())[source]#

Describe which post-fit actions a saved result can reproduce.

Parameters:
summarize_existing_artifactsbool

Whether stored diagnostics can be summarized.

retarget_cached_nuisancesbool

Whether targeting can run again from cached primary nuisances. A guarded DR-TMLE curve can refit reduced regressions inside that step.

evaluate_stored_representerbool

Whether the stored representer can evaluate another parameter.

refit_nuisancesbool

Whether the result retains enough configuration for nuisance refits.

evaluate_new_databool

Whether the fitted result can score new observations.

unreconstructibletuple of str

Missing components that prevent reconstruction.

Methods#

Attributes#

Replayability.unreconstructible: tuple[str, ...] = ()#
Replayability.summarize_existing_artifacts: bool#
Replayability.retarget_cached_nuisances: bool#
Replayability.evaluate_stored_representer: bool#
Replayability.refit_nuisances: bool#
Replayability.evaluate_new_data: bool#