state_arl

ARL (average run length) analyzer for no-reset benchmarks.

class pysatl_cpd.benchmark.online.noreset.tooling.analyzers.state_arl.NoResetArlAnalyzer(entries_picker=None)[source]

Bases: NoResetAnalyzerBase

Analyzer that computes average run length at a range of thresholds.

Parameters:

entries_picker (BenchmarkEntriesPicker | None)

analyze(benchmark_entry, state, thresholds, arl_length)[source]

Compute ARL for each threshold value.

Retrieves no-change runs matching the detector, state, and length, then evaluates the ARL metric at each threshold.

Parameters:

  • benchmark_entry (OnlineNoResetBenchmarkEntry) – Entry describing detector configuration used for picking and validation.

  • state (StateDescriptor) – State descriptor for the no-change runs.

  • thresholds (Sequence[float]) – Threshold values to evaluate.

  • arl_length (int) – Expected length of each no-change run. Must be positive.

Returns:

Table with columns threshold and arl.

Return type:

DataFrame

Raises:

ValueError – If arl_length is not positive.

analyze_runs(runs, thresholds)[source]

Compute ARL for already selected and validated no-change runs.

Return type:

DataFrame

Parameters:
static validate_arl_runs(runs, expected_detector, state, arl_length)[source]

Verify that every run matches the expected detector, state, and length.

Return type:

None

Parameters: