Skip to content

Result Access

MCPipelineResult.test_summaries maps each test step name to an MCResult aggregate.

MCPipelineResult.regression_summaries maps each regression step name to an MCRegressionResult aggregate.

Summary Fields and Methods:

Name Type Description
statistic_trace ndarray Test statistic from each successful replication.
pval_trace ndarray Vectorized p-values for statistic_trace.
mean_statistic float64 Mean test statistic over successful replications.
mean_pval float64 Mean p-value over successful replications.
rejection_rate float64 Share of p-values below alpha.
pval_confidence_interval(...) tuple[float64, float64] Confidence interval for the rejection rate.
statistic_confidence_interval(...) tuple[float64, float64] Confidence interval for the mean test statistic.

If retain_test_results=True, MCPipelineResult.test_results stores scalar per-replication TestResult objects keyed by test step name.

Performance Reporting:

Name Description
report_mc_performance(result) Print the aggregate pipeline throughput report.
report_mc_step_performance(result) Print one throughput report line per pipeline step.
MCPipelineResult.report_performance() Method form of report_mc_performance(...).
MCPipelineResult.report_step_performance() Method form of report_mc_step_performance(...).
Retention and Memory Use

retain_contexts=True and retain_payloads=True can store large arrays from every successful replication. For large Monte Carlo runs, prefer retaining aggregate summaries and scalar test results unless full per-replication payloads are needed for debugging or downstream analysis.