Important

Adds /internal/evaluations/statistics endpoint to fetch evaluation statistics with mean and confidence intervals for specified runs and metrics, updating server and client-side code accordingly.

• Behavior:
  • Adds /internal/evaluations/statistics endpoint in internal.rs to return aggregated statistics (mean, confidence intervals) for specified evaluation runs and metrics.
  • Implements get_evaluation_statistics_handler in get_statistics.rs to handle requests to the new endpoint.
  • Updates TensorZeroClient in tensorzero.ts to include getEvaluationStatistics() method for fetching statistics.
• Data Processing:
  • Adds get_evaluation_statistics method to EvaluationQueries trait in evaluation_queries.rs.
  • Implements ClickHouse query logic in evaluation_queries.rs to fetch raw statistics and compute confidence intervals.
  • Defines EvaluationStatisticsRow struct in evaluation_queries.rs for deserializing statistics data.
• TypeScript Bindings:
  • Adds EvaluationStatistics and GetEvaluationStatisticsResponse types in EvaluationStatistics.ts and GetEvaluationStatisticsResponse.ts.
  • Exports new types in index.ts.
• Testing:
  • Includes unit tests for get_evaluation_statistics in evaluation_queries.rs.
  • Adds E2E tests for the new endpoint in evaluations.rs and evaluation_queries.rs.

^{This description was created by for 9120b40. You can customize this summary. It will automatically update as commits are pushed.}