Struct ConfigReport 

pub struct ConfigReport {
    pub model: String,
    pub n: usize,
    pub accuracy: f64,
    pub cost_total_usd: Option<f64>,
    pub latency_ms_avg: u64,
    pub turns_avg: Option<u64>,
    pub cache_hits: usize,
    pub cache_misses: usize,
    pub flaky_cases: u32,
    pub output_unstable_cases: u32,
    pub by_tag: BTreeMap<String, TagStat>,
    pub cases: Vec<CaseResult>,
}

Fields

model: String

n: usize

accuracy: f64

Mean of each case’s pass_rate. At samples == 1 this is exactly the old passes/n accuracy — the field name is kept for gate continuity.

cost_total_usd: Option<f64>

None when any case’s backend couldn’t report cost — absent is not zero, so a back-end that genuinely can’t supply cost never shows up as “free”.

latency_ms_avg: u64

turns_avg: Option<u64>

Average conversation turns per case (num_turns) — a context-quality diagnostic (fewer turns ⇒ less exploration ⇒ cheaper). 0 in a pre-P4 report. None when any case’s backend couldn’t report a turn count.

cache_hits: usize

Samples (not cases) served from the trace cache (re-scored, 0 LLM) vs freshly re-run this invocation. cache_misses is the count of actual LLM calls made; both 0 in a pre-P4 report.

cache_misses: usize

flaky_cases: u32

Cases with 0 < passes < samples — inconsistent across samples, not a hard regression. 0 in a pre-repeated-sampling report (no case could be flaky at samples == 1).

output_unstable_cases: u32

Cases whose CaseResult::output_stability is Unstable or Incomplete — distinct from flaky_cases, which only tracks verdict flips. A case can be output-unstable while passing (or failing) every sample, which flaky_cases can never see. This is a labelled upper bound, not an exact count — see OutputStability::Unstable for why. 0 when --record-answers was off or samples <= 1 (nothing was measured), same as a pre-repeated-sampling report.

by_tag: BTreeMap<String, TagStat>

cases: Vec<CaseResult>

Trait Implementations

impl Clone for ConfigReport

fn clone(&self) -> ConfigReport

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for ConfigReport

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

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for ConfigReport

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

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl Serialize for ConfigReport

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

where __S: Serializer,

Serialize this value into the given Serde serializer.