Struct RuntimeEvidence 

pub struct RuntimeEvidence {
    pub confirmations: u64,
    pub contradictions: u64,
    pub prior: f64,
    pub last_trace_id: Option<String>,
    pub last_observed_at: DateTime<Utc>,
}

Accumulated runtime-trace evidence for a symbol’s declared effects.

Each asd trace ingest classifies the run as either a confirmation (every observed effect was already declared) or a positive contradiction (an effect was observed at runtime that was NOT declared). A declared effect that simply wasn’t exercised is not a contradiction — absence of observation is not evidence of absence — so it never lowers confidence.

EffectDecl.confidence is then re-derived from these counts via RuntimeEvidence::derive_confidence, seeded by the static prior so that with zero runtime evidence the confidence equals the prior, each confirmation pulls it toward 1.0, and each contradiction toward 0.0.

Fields

confirmations: u64

Traces whose observed effects were all already declared.

contradictions: u64

Traces that observed an effect not present in the declaration.

prior: f64

Static/policy confidence captured at the first runtime observation. Used as the prior the runtime evidence updates, and frozen thereafter so the derivation stays reproducible from (prior, confirmations, contradictions).

last_trace_id: Option<String>

last_observed_at: DateTime<Utc>

Implementations

impl RuntimeEvidence

pub const PRIOR_STRENGTH: f64 = 2.0

Pseudo-count weight given to the static prior. With zero runtime evidence the derived confidence equals the prior; each confirmation pulls it toward 1.0 and each contradiction toward 0.0, with the prior worth PRIOR_STRENGTH observations of resistance.

pub const NEUTRAL_PRIOR: f64 = 0.5

Neutral prior used when no static/policy confidence is available.

pub fn derive_confidence( prior: f64, confirmations: u64, contradictions: u64, ) -> f64

Derive a confidence in [0, 1] from the static prior and accumulated counts. This is a Laplace-smoothed success rate (equivalently the mean of a Beta posterior) with the prior seeded as PRIOR_STRENGTH pseudo-observations:

confidence = (prior*S + confirmations) / (S + confirmations + contradictions)

With confirmations == contradictions == 0 this returns prior exactly.

pub fn confidence(&self) -> f64

Confidence derived from this evidence record.

Trait Implementations

impl Clone for RuntimeEvidence

fn clone(&self) -> Self

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for RuntimeEvidence

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

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for RuntimeEvidence

fn deserialize<__D>(__deserializer: __D) -> Result<Self, __D::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl Serialize for RuntimeEvidence

fn serialize<__S>(&self, __serializer: __S) -> Result<__S::Ok, __S::Error>

where __S: Serializer,

Serialize this value into the given Serde serializer.