pub struct RankingBreakdown {
pub model_id: String,
pub score: Option<f64>,
pub delivery: f64,
pub tool_loop: f64,
pub context: f64,
pub latency: f64,
pub price: f64,
pub weight_delivery: f64,
pub weight_tool_loop: f64,
pub weight_context: f64,
pub weight_latency: f64,
pub weight_price: f64,
}Expand description
Evidence breakdown for one model’s ranking_score - the picker renders
one line per component from this struct, so every ranking decision is
explainable (feedback 02 evidence lines).
The component fields hold the same normalized 0..1 scores used by
ranking_score: the weighted dot product of components × weights IS
the score whenever the score is Some. The picker renders RAW evidence
(success rates, token counts, dollar prices) directly from
ModelVerification/ModelRecord alongside these normalized values.
Fields§
§model_id: String§score: Option<f64>ranking_score result: None when the model is not primary-agent
eligible, unverified, or not passing - unverified never outranks.
delivery: f64§tool_loop: f64§context: f64§latency: f64§price: f64§weight_delivery: f64§weight_tool_loop: f64§weight_context: f64§weight_latency: f64§weight_price: f64Implementations§
Source§impl RankingBreakdown
impl RankingBreakdown
Sourcepub fn breakdown_for(
model: &ModelRecord,
verification: Option<&ModelVerification>,
) -> Self
pub fn breakdown_for( model: &ModelRecord, verification: Option<&ModelVerification>, ) -> Self
Render the full evidence breakdown for one model + verification pair
(normalized component scores, the fixed weights, and the resulting
score). Components are still rendered when the score is None so the
picker can explain WHY the model did not rank.
Trait Implementations§
Source§impl Clone for RankingBreakdown
impl Clone for RankingBreakdown
Source§fn clone(&self) -> RankingBreakdown
fn clone(&self) -> RankingBreakdown
1.0.0 (const: unstable) · Source§fn clone_from(&mut self, source: &Self)
fn clone_from(&mut self, source: &Self)
source. Read more