pub struct BenchResult {Show 15 fields
pub id: i64,
pub bench_run_id: i64,
pub case_id: String,
pub workflow_execution_id: Option<String>,
pub judge_execution_id: Option<String>,
pub score: Option<Value>,
pub headline_score: Option<f64>,
pub status: String,
pub cost_usd: f64,
pub duration_ms: Option<i32>,
pub cache_hit: bool,
pub input_hash: Option<String>,
pub error: Option<String>,
pub created_at: String,
pub workflow_output: Option<Value>,
}Expand description
One per-case score row for a bench run. Carries the workflow execution’s
typed workflow_output alongside the judge’s score blob so the studio’s
typed renderers don’t need a second fetch.
Fields§
§id: i64§bench_run_id: i64§case_id: String§workflow_execution_id: Option<String>§judge_execution_id: Option<String>§score: Option<Value>§headline_score: Option<f64>§status: String§cost_usd: f64§duration_ms: Option<i32>§cache_hit: bool§input_hash: Option<String>§error: Option<String>Human-readable error message captured when status is
workflow_failed or judge_failed; None on ok/cached rows.
Mirrors the server’s BenchResult.error column — present on both the
/bench-runs/{id}/results read path and the live SSE result frame.
created_at: String§workflow_output: Option<Value>Parsed WorkflowEnd payload from the workflow execution. None when
the workflow failed, was canceled, or this is a cache-hit row.
Only the /bench-runs/{id}/results read path (which joins
executions.result) populates this; the live SSE result frame
broadcasts the bare BenchResult row, so this stays None on
events from crate::sub::bench::BenchRunsClient::subscribe_run_events.
Trait Implementations§
Source§impl Clone for BenchResult
impl Clone for BenchResult
Source§fn clone(&self) -> BenchResult
fn clone(&self) -> BenchResult
1.0.0 (const: unstable) · Source§fn clone_from(&mut self, source: &Self)
fn clone_from(&mut self, source: &Self)
source. Read more