Struct ContextPrecision 

pub struct ContextPrecision<M: BaseChatModel> { /* private fields */ }

RAGAS context precision: are the retrieved chunks relevant, and are relevant chunks ranked high?

Each context (in retrieval rank order) is judged relevant to the question; the score is the RAGAS rank-weighted precision:

              K
            ----
         1 \
CP@K = ----- /   v_k · Precision@k
       |REL| ----
            k = 1

where v_k is the binary relevance verdict for the chunk at rank k, Precision@k the share of relevant chunks in the top k, and |REL| the number of relevant chunks. A chunk relevant but buried below irrelevant ones therefore scores lower than the same chunk ranked first.

Implementations

impl<M: BaseChatModel> ContextPrecision<M>

pub fn new(judge: M) -> Self

Creates a context-precision evaluator.

pub fn with_max_context_chars(self, max: usize) -> Self

Per-context character cap sent to the judge.

pub fn with_empty_score(self, score: f64) -> Self

Score when the example carries no retrieved contexts (default 0.0).

Trait Implementations

impl<M: BaseChatModel> RagEvaluator for ContextPrecision<M>

fn eval_rag<'life0, 'life1, 'life2, 'life3, 'life4, 'async_trait>( &'life0 self, input: &'life1 str, _prediction: &'life2 str, contexts: &'life3 [String], _reference: &'life4 str, ) -> Pin<Box<dyn Future<Output = Result<Score, EvalError>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait, 'life1: 'async_trait, 'life2: 'async_trait, 'life3: 'async_trait, 'life4: 'async_trait,

Scores a single prediction given the question, retrieved contexts, and reference answer.

fn name(&self) -> &str

Evaluator name (used in report summaries)