ReferencePython reference

Python reference

Every public name of ergodic 0.2.0, generated from the installed package's signatures and docstrings. One page per module.

Modules#

Module	Contents
Causal graphs	Causal graphs: one mixed-graph core with typed faces (DAG, ADMG, CPDAG, PAG).
Data	Typed datasets: the data objects discovery and inference read.
Domain knowledge	Domain knowledge: prior constraints on causal structure.
Identification	Identification: turn a causal graph into an estimand.
Discovery	Causal discovery: learn causal structure from data.
Effect estimation	Causal inference: estimate causal effects from data and a causal model.
Process intelligence	Process mining and intelligence: from event logs to causal questions.
Structural causal models	The SCM pillar: a Markovian Bayesian structural causal model, PyMC under the hood.
Forecasting	Forecasting: Bayesian trajectories with causal structure in them.
Hidden regimes	Hidden-regime inference: recovering a switch nobody logged.
Datasets	Real-world datasets with their causal question, in the style of sklearn loaders.
Welt bridge	Run Welt's hosted foundation models from ergodic.

Top-level exports#

from ergodic import … reaches these names. Each links to its module page.

GRAPH_FORMAT · IDA · SCM · BinaryGumbelMax · CATEResult · CalibrationReport · CategoricalGumbelMax · CausalForest · CertifyResult · Consensus · Dataset · DecodeResult · DomainKnowledge · Edge · EdgeProbabilities · EffectEstimate · EffectTarget · EventLog · Functional · GaussianAdditive · GraphKind · Hedge · HierarchicalTSData · Hierarchy · IDResult · ITEResult · Identification · InconsistentKnowledgeError · InvalidGraphError · LatentResult · Marginal · Mark · MarkovChain · MediationEstimate · MediationID · MixedGraph · Ordinal · PanelData · PerUnitEffect · PolicyTree · Prob · Product · Ratio · Refutation · RegimeLikelihood · SCMResult · Schema · SchemaInferenceWarning · SelectionReport · Sensitivity · TabularData · TimeSeriesData · VariableImportance · VariableType · adjustment_set · admg · backtest · calibrate · case_table · cate · certify · conformance · consensus · cpdag · d_separated · dag · datasets · decision_points · decision_table · decode · did · discover · discover_model · estimate · estimate_effect · estimate_mediation · event_log · event_study · forecast · frontdoor_set · granger · hts · ida · identify_distribution · identify_effect · identify_mediation · instruments · interrupted_time_series · is_frontdoor_set · is_instrument · is_valid_adjustment_set · kpi_panel · latent_model · m_separated · minimal_adjustment_set · mixed · pag · panel · policy_tree · process_map · recover · refute_effect · select_cate · synthetic_control · tabular · timeseries

GRAPH_FORMAT#

python

GRAPH_FORMAT: str = 'ergodic.graph/1'

The serialization format to_dict() writes and from_dict() reads.

A dict without a "format" field (written before the field existed, or by hand) is read as this version. Any other value is refused, so a reader never guesses at a layout it does not know.

Suggest an edit