pub struct CaseRun {
pub case: String,
pub skill: String,
pub platform: String,
pub model: String,
pub passed: bool,
pub turns: usize,
pub evals: Vec<EvalOutcome>,
pub transcript: Transcript,
pub usage: Option<Usage>,
pub mock_calls: Option<Vec<MockCall>>,
pub history_command: Option<String>,
}Expand description
The result of running one test case on one (platform, model) pair.
Fields§
§case: StringThe test case name.
skill: StringAbsolute-ish path to the skill that was exercised.
platform: StringThe harness platform this run used.
model: StringThe model this run used.
passed: boolTrue iff every eval in this run passed.
turns: usizeNumber of assistant turns produced.
evals: Vec<EvalOutcome>Per-eval outcomes, in declaration order.
transcript: TranscriptThe full conversation, for debugging and deterministic mix-in checks.
usage: Option<Usage>Aggregated token/cost usage across every provider call in this run (skill turns + simulated-user turns + judge calls). Omitted when no usage was reported (e.g. the fake provider or a harness that doesn’t surface usage).
mock_calls: Option<Vec<MockCall>>Every tool call the mock/spy channel observed, in order, with the
original (pre-rewrite) input and the verdict applied. null when the
channel was off for this run (no mocks, no spy); an empty array
means the channel was on and the skill made no tool calls — SDKs use
that distinction so a spy on a channel-less run errs instead of reading
as “zero calls”.
history_command: Option<String>A ready-to-run command that replays this run’s recorded transcript — e.g.
oneharness history show <name> --history-dir <dir> — so a past run can
be reviewed after the fact. Present only when the run was recorded (the
oneharness provider with history enabled); null for providers/configs
that record no history.
Trait Implementations§
Source§impl<'de> Deserialize<'de> for CaseRun
impl<'de> Deserialize<'de> for CaseRun
Source§fn deserialize<__D>(__deserializer: __D) -> Result<Self, __D::Error>where
__D: Deserializer<'de>,
fn deserialize<__D>(__deserializer: __D) -> Result<Self, __D::Error>where
__D: Deserializer<'de>,
Source§impl JsonSchema for CaseRun
impl JsonSchema for CaseRun
Source§fn schema_id() -> Cow<'static, str>
fn schema_id() -> Cow<'static, str>
Source§fn json_schema(generator: &mut SchemaGenerator) -> Schema
fn json_schema(generator: &mut SchemaGenerator) -> Schema
Source§fn inline_schema() -> bool
fn inline_schema() -> bool
$ref keyword. Read more