Struct TaskCompletionEvaluator 

pub struct TaskCompletionEvaluator { /* private fields */ }

Task completion against a declared assertion (prompt: task_completion_v0).

Criterion: the case must declare an expected_assertion and the invocation must have a non-empty final response.

Implementations

impl TaskCompletionEvaluator

pub const fn new(config: JudgeEvaluatorConfig) -> Self

Construct with the supplied judge config.

pub fn with_prompt(self, template: Arc<dyn JudgePromptTemplate>) -> Self

Override the prompt template used by this evaluator.

pub fn with_few_shot(self, examples: Vec<FewShotExample>) -> Self

Attach evaluator-level few-shot examples that render before any case-level examples.

pub fn with_system_prompt(self, prompt: impl Into<String>) -> Self

Override the system prompt visible to the template render.

pub fn with_output_schema(self, schema: Value) -> Self

Attach an output schema for custom prompt templates.

pub fn with_use_reasoning(self, flag: bool) -> Self

Toggle judge reasoning capture.

pub fn with_feedback_key(self, key: impl Into<String>) -> Self

Override the feedback key used by downstream exporters.

pub const fn config(&self) -> &JudgeEvaluatorConfig

Borrow the underlying config (e.g., to inspect the judge registry or feedback key).

Trait Implementations

impl Evaluator for TaskCompletionEvaluator

fn name(&self) -> &'static str

Unique name for this evaluator, used in EvalCase::evaluators filters.

fn evaluate( &self, case: &EvalCase, invocation: &Invocation, ) -> Option<EvalMetricResult>

Score the actual invocation against the expected case.

impl JudgeEvaluatorBuilder for TaskCompletionEvaluator

fn judge_config_mut(&mut self) -> &mut JudgeEvaluatorConfig

Borrow the evaluator’s underlying JudgeEvaluatorConfig for mutation by the default builder methods.

fn with_prompt(self, template: Arc<dyn JudgePromptTemplate>) -> Self

Override the built-in prompt template.

fn with_few_shot(self, examples: Vec<FewShotExample>) -> Self

Attach few-shot examples.

fn with_system_prompt(self, prompt: impl Into<String>) -> Self

Override the system prompt applied ahead of the rendered prompt.

fn with_output_schema(self, schema: Value) -> Self

Override the output-schema identifier used by structured-output evaluators.

fn with_use_reasoning(self, flag: bool) -> Self

Toggle the use_reasoning flag.

fn with_feedback_key(self, key: impl Into<String>) -> Self

Override the feedback key forwarded to telemetry / reporter backends.

fn with_aggregator(self, aggregator: Arc<dyn Aggregator>) -> Self

Override the aggregator applied to per-sample judge scores.