feat: Add metadata to evaluation result.

PiperOrigin-RevId: 660589553

Author: jsondai

vertexai/preview/evaluation/_base.py

@@ -37,7 +37,7 @@ class EvaluationRunConfig:
 
     Attributes:
       dataset: The dataset to evaluate.
-      metrics: The list of metric names, or metric bundle names, or Metric instances to evaluate.
+      metrics: The list of metrics, or Metric instances to evaluate.
       column_map: The dictionary of column name overrides in the dataset.
       client: The evaluation service client.
       evaluation_service_qps: The custom QPS limit for the evaluation service.
@@ -73,10 +73,11 @@ class EvalResult:
     """Evaluation result.
 
     Attributes:
-      summary_metrics: The summary evaluation metrics for an evaluation run.
-      metrics_table: A table containing eval inputs, ground truth, and metrics per
-        row.
+      summary_metrics: The summary evaluation metrics for the evaluation run.
+      metrics_table: A table containing evaluation dataset, and metric results.
+      metadata: The metadata for the evaluation run.
     """
 
     summary_metrics: Dict[str, float]
     metrics_table: Optional["pd.DataFrame"] = None
+    metadata: Optional[Dict[str, str]] = None

vertexai/preview/evaluation/_eval_tasks.py

@@ -298,6 +298,10 @@ def _evaluate_with_experiment(
                 k: ("NaN" if isinstance(v, float) and np.isnan(v) else v)
                 for k, v in eval_result.summary_metrics.items()
             }
+            eval_result.metadata = {
+                "experiment": self.experiment,
+                "experiment_run": experiment_run_name,
+            }
             try:
                 vertexai.preview.log_metrics(eval_result.summary_metrics)
             except (TypeError, exceptions.InvalidArgument) as e: