API reference

One page per public class and function, generated from docstrings via sphinx.ext.autosummary. Each ## section below is a real import path (xcausal, xcausal.aggregation, …); each item is listed once, under the module it’s defined in.

Many aggregation-strategy and adapter classes (e.g. Median, CausalMLAdapter) are re-exported at the top level for convenience – from xcausal import Median works even though it’s documented under xcausal.aggregation below. See the Mixed-framework method list or xcausal.__all__ for the exact top-level surface.

xcausal

Top-level namespace: the ensemble class, its estimate/result types, and outcome-type detection.

xcausal.CausalEnsemble	Ensemble of causal estimators with robust aggregation.
xcausal.AteEstimate	Result from the ensemble of causal estimators.
xcausal.CateEstimate	Ensemble CATE result.
xcausal.ComponentAteEstimate	Result from a single causal estimator.
xcausal.ComponentCateEstimate	CATE result from a single component method.
xcausal.infer_outcome_type	Detect the outcome type from the values of Y.

Warnings

xcausal.XCausalWarning	Umbrella class for all xcausal-emitted warnings.
xcausal.ComponentWarning	Base class for warnings indicating an operating-ensemble pool reduction.
xcausal.ComponentFailureWarning	A component raised an exception during fit, cross-fitting, prediction, or as part of nuisance/weight-fitting machinery, and was silently excluded from the operating ensemble.
xcausal.ComponentExclusionWarning	A component was excluded by configuration mismatch.
xcausal.BootstrapWarning	A bootstrap inference call hit a reliability concern.
xcausal.RiskDifferenceWarning	A binary-outcome ensemble point estimate is outside [-1, 1].

xcausal.defaults

Factories for the shipped component pools and their outcome/propensity models. See Default components and model configuration for generated rosters, resolved settings, and independent member effect-model examples. Supervised aggregation’s nuisance arguments configure its weight-learning loss; member models are configured through CausalEnsemble(methods=[...]).

xcausal.defaults.default_methods	Return a default set of causal estimators for the given outcome type.
xcausal.defaults.default_outcome_model	Default outcome model for the given outcome type.
xcausal.defaults.default_propensity_model	Default propensity model: HistGradientBoostingClassifier.

xcausal.aggregation

Aggregation strategies (combine component ATE/CATE predictions into an ensemble estimate) and the result types they produce.

xcausal.aggregation.AggregationStrategy	Abstract base class for all aggregation strategies.
xcausal.aggregation.PointwiseStrategy	Base class for strategies that apply a fixed statistical rule pointwise.
xcausal.aggregation.AgreementStrategy	Base class for strategies that select/weight models by inter-model agreement.
xcausal.aggregation.SupervisedStrategy	Base class for outcome-supervised aggregation strategies.
xcausal.aggregation.Mean	Pointwise mean aggregation.
xcausal.aggregation.Median	Pointwise median aggregation.
xcausal.aggregation.TrimmedMean	Pointwise trimmed-mean aggregation.
xcausal.aggregation.CBA	Consensus Based Averaging (Machluf et al., 2024).
xcausal.aggregation.CausalStacking	Causal Stacking (Han & Wu, 2022).
xcausal.aggregation.QAggregation	Doubly Robust Q-Aggregation (Lan & Syrgkanis, 2024).
xcausal.aggregation.RStacking	R-Stacking (Nie & Wager, 2021).
xcausal.aggregation.Select	Select the single component model minimizing a pseudo-outcome risk.
xcausal.aggregation.EnsembleWeights	Result of weight computation for any non-pointwise aggregation.
xcausal.aggregation.BootstrapResult	Bootstrap inference results for both ATE and CATE.

Cross-fitting

xcausal.aggregation.CrossFitSplit	Q-fold cross-fitting split.
xcausal.aggregation.TrainAvgSplit	Simple train / averaging-set split.
xcausal.aggregation.FoldSpec	Result of a data split: absolute indices into the original array per fold.

Nuisance estimation

xcausal.aggregation.NuisanceEstimates	Out-of-fold nuisance model predictions.
xcausal.aggregation.fit_nuisance	Fit nuisance models via cross-fitting and return out-of-fold predictions.
xcausal.aggregation.dr_pseudo_outcome	Compute the DR/AIPW pseudo-outcome for each observation.
xcausal.aggregation.robinson_residuals	Compute Robinson decomposition residuals.

xcausal.adapters

Wrappers presenting each supported library’s estimators (EconML, CausalML, DoubleML, stochtree) through one common interface, plus generic adapters for wrapping arbitrary user-supplied estimators.

xcausal.adapters.CausalEstimator	Protocol that all causal estimator adapters must satisfy.
xcausal.adapters.EconMLAdapter	Wrap an initialized-but-unfitted EconML estimator.
xcausal.adapters.CausalMLAdapter	Wrap a CausalML estimator as a xcausal causal estimator.
xcausal.adapters.DoubleMLAdapter	Wrap a DoubleML estimator class for use in CausalEnsemble.
xcausal.adapters.StochtreeAdapter	Wrap stochtree's BCFModel for use in CausalEnsemble.
xcausal.adapters.GenericATEAdapter	Wrap a callable fn(X, T, Y) -> float as an ATE-only causal estimator.
xcausal.adapters.GenericCATEAdapter	Wrap callables as a CATE-capable causal estimator.
xcausal.adapters.INNER_WORKER_ENV	Sentinel env var set inside every XCausal parallel worker.

xcausal.plots

Requires the plots extra (pip install "xcausal[plots]"). Each function is also available as a thin method on the class it plots – see the corresponding class above (e.g. BootstrapResult.forest()).

xcausal.plots.forest	Forest plot of per-component and ensemble ATEs.
xcausal.plots.weights	Bar chart of aggregation weights.
xcausal.plots.cate_profile	Ensemble CATE along one covariate, with optional bootstrap CI.
xcausal.plots.disagreement	Pairwise disagreement between component CATEs evaluated on X.

xcausal.datasets

xcausal.datasets.load_lalonde

Load the Lalonde job training dataset.