Feat: Agent Evaluations print detailed results when all tests pass

contributing/samples/gepa/experiment.py

@@ -43,7 +43,6 @@
 from tau_bench.types import EnvRunResult
 from tau_bench.types import RunConfig
 import tau_bench_agent as tau_bench_agent_lib
-
 import utils
 
 

contributing/samples/gepa/run_experiment.py

@@ -25,7 +25,6 @@
 from absl import flags
 import experiment
 from google.genai import types
-
 import utils
 
 _OUTPUT_DIR = flags.DEFINE_string(

src/google/adk/cli/cli_tools_click.py

@@ -561,11 +561,22 @@ def wrapper(*args, **kwargs):
     help="Optional. Whether to print detailed results on console or not.",
 )
 @eval_options()
+@click.option(
+    "--print_detailed_results_on_success",
+    is_flag=True,
+    show_default=True,
+    default=False,
+    help=(
+        "Optional. Whether to print detailed results on console or not even if"
+        " the evaluation passed."
+    ),
+)
 def cli_eval(
     agent_module_file_path: str,
     eval_set_file_path_or_id: list[str],
     config_file_path: str,
     print_detailed_results: bool,
+    print_detailed_results_on_success: bool,
     eval_storage_uri: Optional[str] = None,
     log_level: str = "INFO",
 ):
@@ -780,6 +791,11 @@ def cli_eval(
   if print_detailed_results:
     for eval_result in eval_results:
       eval_result: EvalCaseResult
+
+    if (
+        eval_result.final_eval_status != EvalStatus.PASSED
+        or print_detailed_results_on_success
+    ):
       click.echo(
           "********************************************************************"
       )

src/google/adk/evaluation/agent_evaluator.py

@@ -113,6 +113,7 @@ async def evaluate_eval_set(
       num_runs: int = NUM_RUNS,
       agent_name: Optional[str] = None,
       print_detailed_results: bool = True,
+      print_detailed_results_on_success: bool = False,
   ):
     """Evaluates an agent using the given EvalSet.
 
@@ -178,6 +179,7 @@ async def evaluate_eval_set(
       failures_per_eval_case = AgentEvaluator._process_metrics_and_get_failures(
           eval_metric_results=eval_metric_results,
           print_detailed_results=print_detailed_results,
+          print_detailed_results_on_success=print_detailed_results_on_success,
           agent_module=agent_name,
       )
 
@@ -200,6 +202,7 @@ async def evaluate(
       agent_name: Optional[str] = None,
       initial_session_file: Optional[str] = None,
       print_detailed_results: bool = True,
+      print_detailed_results_on_success: bool = False,
   ):
     """Evaluates an Agent given eval data.
 
@@ -245,6 +248,7 @@ async def evaluate(
           num_runs=num_runs,
           agent_name=agent_name,
           print_detailed_results=print_detailed_results,
+          print_detailed_results_on_success=print_detailed_results_on_success,
       )
 
   @staticmethod
@@ -648,6 +652,7 @@ def _get_eval_metric_results_with_invocation(
   def _process_metrics_and_get_failures(
       eval_metric_results: dict[str, list[_EvalMetricResultWithInvocation]],
       print_detailed_results: bool,
+      print_detailed_results_on_success: bool,
       agent_module: str,
   ) -> list[str]:
     """Returns a list of failures based on the score for each invocation."""
@@ -678,17 +683,25 @@ def _process_metrics_and_get_failures(
 
       # Gather all the failures.
       if overall_eval_status != EvalStatus.PASSED:
-        if print_detailed_results:
-          AgentEvaluator._print_details(
-              eval_metric_result_with_invocations=eval_metric_results_with_invocations,
-              overall_eval_status=overall_eval_status,
-              overall_score=overall_score,
-              metric_name=metric_name,
-              threshold=threshold,
-          )
         failures.append(
             f"{metric_name} for {agent_module} Failed. Expected {threshold},"
             f" but got {overall_score}."
         )
 
+      should_print = print_detailed_results and (
+          overall_eval_status != EvalStatus.PASSED
+          or print_detailed_results_on_success
+      )
+
+      if should_print:
+        AgentEvaluator._print_details(
+            eval_metric_result_with_invocations=(
+                eval_metric_results_with_invocations
+            ),
+            overall_eval_status=overall_eval_status,
+            overall_score=overall_score,
+            metric_name=metric_name,
+            threshold=threshold,
+        )
+
     return failures