Struct ExtractionStatsBuilder 

pub struct ExtractionStatsBuilder<'a> { /* private fields */ }

Per-call builder returned by Client::extraction_stats.

Awaiting the builder computes extraction accuracy per (provider, model) over the requested scope slice — a read-only Postgres aggregate, no LLM call. Each ExtractionStat row carries the total semantic rows produced and the subset retired as rejected (a wrong extraction the user corrected); ExtractionStat::accuracy is the derived ratio.

The scope setters narrow the slice and are AND-combined; an unset dimension imposes no constraint, so the default (no setters) aggregates the whole store. Use .org(..) for a per-tenant number, add .agent(..) / .user(..) to narrow further.

Examples

// Per-model accuracy for one org.
for stat in client.extraction_stats().org("acme").await? {
    println!("{}/{}: {:.1}% over {} rows", stat.provider, stat.model, stat.accuracy() * 100.0, stat.total);
}

Implementations

impl<'a> ExtractionStatsBuilder<'a>

pub fn agent(self, agent_id: impl Into<String>) -> Self

Narrows the slice to one agent id. Default: all agents.

pub fn org(self, org_id: impl Into<String>) -> Self

Narrows the slice to one org id. Default: all orgs.

pub fn user(self, user_id: impl Into<String>) -> Self

Narrows the slice to one user id. Default: all users.

Trait Implementations

impl<'a> IntoFuture for ExtractionStatsBuilder<'a>

type Output = Result<Vec<ExtractionStat>, ClientError>

The output that the future will produce on completion.

type IntoFuture = Pin<Box<dyn Future<Output = <ExtractionStatsBuilder<'a> as IntoFuture>::Output> + Send + 'a>>

Which kind of future are we turning this into?

fn into_future(self) -> Self::IntoFuture

Creates a future from a value.