Struct EvalContext 

pub struct EvalContext {

    pub session_id: String,
    pub run_id: Option<String>,
    pub iteration: Option<u32>,
    pub input_tokens: Option<u64>,
    pub output_tokens: Option<u64>,
    pub tokens_remaining: Option<u64>,
    pub total_tokens_used: Option<u64>,
    pub tool_call_count: Option<u32>,
    pub tool_error_count: Option<u32>,
    pub tool_name: Option<String>,
    pub tool_errored: Option<bool>,
    pub max_iterations: Option<u32>,
    pub metadata: HashMap<String, String>,
}

Context provided to evaluators for scoring.

Carries the information an evaluator needs without requiring it to depend on Arcan types directly.

Fields

session_id: String

Session ID.

run_id: Option<String>

Run ID within the session.

iteration: Option<u32>

Current iteration within the run.

input_tokens: Option<u64>

Input token count for the current model call.

output_tokens: Option<u64>

Output token count for the current model call.

tokens_remaining: Option<u64>

Remaining token budget.

total_tokens_used: Option<u64>

Total tokens used so far in the session.

tool_call_count: Option<u32>

Number of tool calls in this run.

tool_error_count: Option<u32>

Number of tool errors in this run.

tool_name: Option<String>

Tool name (for tool-specific evaluators).

tool_errored: Option<bool>

Whether the tool call resulted in an error.

max_iterations: Option<u32>

Maximum iterations configured.

metadata: HashMap<String, String>

Arbitrary key-value metadata.

Implementations

impl EvalContext

pub fn new(session_id: impl Into<String>) -> Self

Create a minimal context with just a session ID.

Trait Implementations

impl Clone for EvalContext

fn clone(&self) -> EvalContext

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for EvalContext

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

Formats the value using the given formatter.