Struct ScoringPolicy 

pub struct ScoringPolicy {
    pub schema_version: u32,
    pub attribution: AttributionMultipliers,
    pub verification: VerificationMultipliers,
    pub freshness: FreshnessParams,
    pub deprecation: DeprecationParams,
    pub version_match: VersionMatchMultipliers,
    pub blend: BlendWeights,
}

Full scoring-policy shape — every section is independently overridable.

Fields

schema_version: u32

Schema sentinel. Always 1 in v1.

attribution: AttributionMultipliers

Attribution-based trust multipliers.

verification: VerificationMultipliers

Verification-based trust multipliers.

freshness: FreshnessParams

Freshness decay parameters.

deprecation: DeprecationParams

Deprecation penalty.

version_match: VersionMatchMultipliers

Version-target match multipliers.

blend: BlendWeights

Trust × relevance blend weights.

Implementations

impl ScoringPolicy

pub const fn attribution_multiplier(&self, attribution: Attribution) -> f64

The attribution multiplier for one Attribution variant.

pub fn verification_multiplier( &self, verified: bool, verified_by: Option<&str>, ) -> f64

The verification multiplier, keyed off whether the content was verified and who verified it. Unverified content always lands on the lowest unverified multiplier, so any verified result outranks it (#3).

pub fn freshness_multiplier(&self, age_days: i64) -> f64

The exponential freshness multiplier exp(-age_days / half_life_days). Negative ages (clock skew) are treated as zero (fully fresh).

pub fn version_multiplier(&self, class: &MatchClass) -> f64

Map a crate::version_match::MatchClass to its trust multiplier using the [version_match] policy knobs (spec §3.3): linear steps with a floor. Breaking maps to floor for completeness — callers drop Breaking candidates before scoring.

pub fn confidence(&self, trust_score: f64, relevance: f64) -> f64

Blend trust and relevance into a confidence in [0, 1].

confidence = clamp(trust^trust_weight * relevance^relevance_weight).

pub fn score( &self, provenance: &Provenance, version: Option<&VersionScoreInput>, age_days: i64, relevance: f64, relevance_source: RelevanceSource, ) -> ScoreResult

Score one candidate end to end: trust, blended confidence, and the per-factor breakdown. relevance must already be normalized to [0, 1] — via normalize_rrf for RRF scores, or directly for Voyage relevance scores (which Voyage already returns in [0, 1]).

impl ScoringPolicy

pub fn parse(body: &str) -> Result<Self, ScoringPolicyError>

Parse a scoring-policy TOML body.

Errors

Returns ScoringPolicyError::Parse if the TOML is malformed, ScoringPolicyError::SchemaVersionMismatch if the schema sentinel disagrees, or ScoringPolicyError::NonFiniteWeight if any numeric weight is not a finite, non-negative f64.

Trait Implementations

impl Clone for ScoringPolicy

fn clone(&self) -> ScoringPolicy

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for ScoringPolicy

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

Formats the value using the given formatter.

impl Default for ScoringPolicy

fn default() -> Self

Returns the “default value” for a type.

impl<'de> Deserialize<'de> for ScoringPolicy

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

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl PartialEq for ScoringPolicy

fn eq(&self, other: &ScoringPolicy) -> bool

Tests for self and other values to be equal, and is used by ==.

fn ne(&self, other: &Rhs) -> bool

Tests for !=. The default implementation is almost always sufficient, and should not be overridden without very good reason.

impl Serialize for ScoringPolicy

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

where __S: Serializer,

Serialize this value into the given Serde serializer.

impl StructuralPartialEq for ScoringPolicy