/** * ThinkHive SDK - Eval Runs API * * Evaluation run management for agent quality assessment */ /** Options for creating an evaluation run */ export interface CreateEvalRunOptions { traceIds?: string[]; criteriaIds?: string[]; confidenceLevel?: number; useLLM?: boolean; } /** Options for listing evaluation run results */ export interface GetEvalRunResultsOptions { limit?: number; offset?: number; passedOnly?: boolean; failedOnly?: boolean; } /** Options for listing evaluation runs */ export interface ListEvalRunsOptions { agentId?: string; limit?: number; } /** Options for estimating evaluation cost */ export interface EstimateCostOptions { traceIds?: string[]; criteriaIds?: string[]; confidenceLevel?: number; useLLM?: boolean; } /** Options for getting trace-level results */ export interface GetTraceResultsOptions { latest?: boolean; includeCriteria?: boolean; } /** An evaluation run */ export interface EvalRun { id: string; agentId: string; status: string; traceCount: number; criteriaCount: number; createdAt: string; completedAt?: string; } /** A single evaluation result */ export interface EvalResult { id: string; runId: string; traceId: string; criterionId: string; passed: boolean; score: number; reasoning?: string; } /** Cost estimate for an evaluation run */ export interface EvalCostEstimate { estimatedTraces: number; estimatedCriteria: number; estimatedCost: number; estimatedCredits: number; } /** * Evaluation runs API client for managing agent evaluation workflows */ export declare const evalRuns: { /** * Create a new evaluation run for an agent * * @param agentId - The agent to evaluate * @param opts - Optional configuration for the evaluation run * @returns The created evaluation run */ create(agentId: string, opts?: CreateEvalRunOptions): Promise; /** * Get an evaluation run by ID * * @param runId - The evaluation run ID * @returns The evaluation run details */ get(runId: string): Promise; /** * Get results for an evaluation run * * @param runId - The evaluation run ID * @param opts - Pagination and filter options * @returns Paginated evaluation results */ getResults(runId: string, opts?: GetEvalRunResultsOptions): Promise<{ results: EvalResult[]; limit: number; offset: number; hasMore: boolean; }>; /** * List evaluation runs with optional filters * * @param opts - Filter and pagination options * @returns List of evaluation runs */ list(opts?: ListEvalRunsOptions): Promise; /** * Estimate the cost of running an evaluation * * @param agentId - The agent to estimate cost for * @param opts - Evaluation configuration for cost estimation * @returns Cost estimate details */ estimateCost(agentId: string, opts?: EstimateCostOptions): Promise; /** * Get evaluation results for a specific trace * * @param traceId - The trace ID to get results for * @param opts - Options for filtering results * @returns Evaluation results for the trace */ getTraceResults(traceId: string, opts?: GetTraceResultsOptions): Promise; /** * Delete an evaluation run * * @param runId - The evaluation run ID to delete */ remove(runId: string): Promise; }; export { evalRuns as default };