`MultiSampleAnalysis.detect_events`¶

Module: workflow.py

Estimate a baseline and detect events on every loaded/denoised trace. For local parameter tuning, use detect_events_simple first.

Parameters¶

detect_method (str): threshold, zscore_threshold, cusum, pelt, or hmm; default "threshold".
detect_params (dict[str, Any] | None): Method-specific parameters. If omitted, method defaults are used.
baseline_method (str): "rolling_quantile" by default; also supports "global_quantile" and "global_median".
baseline_params (dict[str, Any] | None): Workflow fallback is {"window": 10000, "q": 0.5}. For global_quantile, use {"q": ...}; for global_median, no parameter is required.
detect_direction (str): "down" by default or "up".
merge_event (bool): Merge nearby events when True; default False.
merge_event_params (dict[str, Any] | None): Use {"merge_gap_ms": ...} when merging.
exclude_current (bool): Use a current mask for baseline/noise statistics; default True.
exclude_current_params (dict[str, Any] | None): Bounds in the form {"min": ..., "max": ...}. When omitted, defaults are direction-dependent (up: below 0; down: above 0).

MultiSampleAnalysis. Events are stored in analysis.events, baselines in analysis.baselines, and settings in analysis.detect_state.

A tqdm progress bar is used when available.
At most one effective point after current masking raises ValueError.
global_quantile is the Agent Skill policy default, but the Python method default remains rolling_quantile.