Enum AdaptiveOutcome 

pub enum AdaptiveOutcome {
    LeakDetected {
        posterior: Posterior,
        samples_per_class: usize,
        elapsed_secs: f64,
    },
    NoLeakDetected {
        posterior: Posterior,
        samples_per_class: usize,
        elapsed_secs: f64,
    },
    Inconclusive {
        reason: InconclusiveReason,
        posterior: Option<Posterior>,
        samples_per_class: usize,
        elapsed_secs: f64,
    },
    ThresholdElevated {
        posterior: Posterior,
        theta_user: f64,
        theta_eff: f64,
        theta_tick: f64,
        achievable_at_max: bool,
        samples_per_class: usize,
        elapsed_secs: f64,
    },
}

Outcome of an adaptive sampling decision (no_std compatible).

Uses f64 for elapsed time instead of std::time::Duration.

Variants

LeakDetected

Leak probability exceeded fail threshold - timing leak detected.

Fields

posterior: Posterior

Final posterior distribution.

samples_per_class: usize

Number of samples collected per class.

elapsed_secs: f64

Time spent in seconds.

NoLeakDetected

Leak probability dropped below pass threshold - no significant leak.

Fields

posterior: Posterior

Final posterior distribution.

samples_per_class: usize

Number of samples collected per class.

elapsed_secs: f64

Time spent in seconds.

Inconclusive

A quality gate triggered before reaching a decision.

Fields

reason: InconclusiveReason

Reason for stopping.

posterior: Option<Posterior>

Final posterior distribution (if available).

samples_per_class: usize

Number of samples collected per class.

elapsed_secs: f64

Time spent in seconds.

ThresholdElevated

Threshold was elevated and pass criterion was met at effective threshold (v5.5).

P < pass_threshold at θ_eff, but θ_eff > θ_user + ε. This is semantically distinct from both Pass (can’t guarantee user threshold) and Inconclusive (not a quality issue - we got a clear statistical result).

Fields

posterior: Posterior

Final posterior distribution.

theta_user: f64

User’s requested threshold (θ_user).

theta_eff: f64

Effective threshold used (θ_eff).

theta_tick: f64

Timer tick floor (θ_tick).

achievable_at_max: bool

Whether threshold is achievable at max_samples.

samples_per_class: usize

Number of samples collected per class.

elapsed_secs: f64

Time spent in seconds.

Implementations

impl AdaptiveOutcome

pub fn leak_probability(&self) -> Option<f64>

Get the final leak probability, if available.

pub fn is_leak_detected(&self) -> bool

Check if the outcome indicates a leak was detected.

pub fn is_conclusive(&self) -> bool

Check if the outcome is conclusive (either pass or fail).

Note: ThresholdElevated is NOT considered conclusive in v5.5 - it means we got a statistical result but can’t guarantee the user’s threshold.

pub fn is_threshold_elevated(&self) -> bool

Check if the threshold was elevated beyond tolerance (v5.5).

pub fn samples_per_class(&self) -> usize

Get the number of samples per class.

pub fn elapsed_secs(&self) -> f64

Get elapsed time in seconds.

pub fn to_summary(&self, calibration: &Calibration) -> OutcomeSummary

Convert to an FFI-friendly summary containing only scalar fields.

This centralizes the conversion logic that was previously duplicated across NAPI and C bindings, using canonical effect pattern classification.

Trait Implementations

impl Clone for AdaptiveOutcome

fn clone(&self) -> AdaptiveOutcome

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for AdaptiveOutcome

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.