Struct TraceExpects 

pub struct TraceExpects {
    pub response_contains: Vec<String>,
    pub response_not_contains: Vec<String>,
    pub tools_used: Vec<String>,
    pub tools_not_used: Vec<String>,
    pub max_tool_calls: Option<usize>,
    pub all_tools_succeeded: Option<bool>,
    pub response_matches: Vec<String>,
    pub min_tool_calls: Option<usize>,
    pub exact_tool_calls: Option<usize>,
    pub tool_arguments_contain: Vec<ToolPayloadExpect>,
    pub tool_results_contain: Vec<ToolPayloadExpect>,
}

Declarative expectations for grading a run.

Unknown keys are rejected so a typoed expectation cannot degrade a case into a silent no-op. See TraceExpects::is_empty for the companion guard against a fixture that declares no expectation at all.

Fields

response_contains: Vec<String>

Substrings the final response must contain.

response_not_contains: Vec<String>

Substrings the final response must NOT contain.

tools_used: Vec<String>

Tool names that must have been called.

tools_not_used: Vec<String>

Tool names that must NOT have been called.

max_tool_calls: Option<usize>

Upper bound on the number of tool calls.

all_tools_succeeded: Option<bool>

If set, whether every tool call must have succeeded.

response_matches: Vec<String>

Regex patterns the final response must match.

min_tool_calls: Option<usize>

Lower bound on the number of tool calls.

exact_tool_calls: Option<usize>

Exact number of tool calls the run must have made. Unlike tools_used (existential) and max_tool_calls (upper bound), this fails when a dispatch the fixture claims is missing.

tool_arguments_contain: Vec<ToolPayloadExpect>

Substrings that must appear in the arguments dispatched to a tool.

tool_results_contain: Vec<ToolPayloadExpect>

Substrings that must appear in the result a tool returned.

Implementations

impl TraceExpects

pub fn is_empty(&self) -> bool

True when the fixture declares no effective assertion.

evaluate_expects produces one grade per declared expectation, and CaseReport::passed() is vacuously true over zero grades. A case in this state exercises the agent but certifies nothing, so a required gate must reject it at load time rather than report it green.

Trait Implementations

impl Clone for TraceExpects

fn clone(&self) -> Self

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for TraceExpects

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

Formats the value using the given formatter.

impl Default for TraceExpects

fn default() -> Self

Returns the “default value” for a type.

impl<'de> Deserialize<'de> for TraceExpects

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

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl Serialize for TraceExpects

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

where __S: Serializer,

Serialize this value into the given Serde serializer.