Define EvalReport data model.

PiperOrigin-RevId: 757933585

src/google/adk/cli/cli_eval.py

@@ -12,6 +12,7 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
+import datetime
 from enum import Enum
 import importlib.util
 import json
@@ -25,8 +26,10 @@ from typing import Optional
 import uuid
 
 from pydantic import BaseModel
+from pydantic import Field
 
 from ..agents import Agent
+from ..sessions.session import Session
 
 logger = logging.getLogger(__name__)
 
@@ -43,16 +46,25 @@ class EvalMetric(BaseModel):
 
 
 class EvalMetricResult(BaseModel):
-  score: Optional[float]
+  score: Optional[float] = None
   eval_status: EvalStatus
 
 
-class EvalResult(BaseModel):
+class EvalCaseResult(BaseModel):
   eval_set_file: str
   eval_id: str
   final_eval_status: EvalStatus
   eval_metric_results: list[tuple[EvalMetric, EvalMetricResult]]
   session_id: str
+  session_details: Optional[Session] = None
+
+
+class EvalSetResult(BaseModel):
+  eval_set_result_id: str
+  eval_set_result_name: str
+  eval_set_id: str
+  eval_case_results: list[EvalCaseResult] = Field(default_factory=list)
+  creation_timestamp: float = 0.0
 
 
 MISSING_EVAL_DEPENDENCIES_MESSAGE = (
@@ -154,7 +166,7 @@ async def run_evals(
     session_service=None,
     artifact_service=None,
     print_detailed_results=False,
-) -> AsyncGenerator[EvalResult, None]:
+) -> AsyncGenerator[EvalCaseResult, None]:
   try:
     from ..evaluation.agent_evaluator import EvaluationGenerator
     from ..evaluation.response_evaluator import ResponseEvaluator
@@ -249,7 +261,7 @@ async def run_evals(
           else:
             raise ValueError("Unknown eval status.")
 
-        yield EvalResult(
+        yield EvalCaseResult(
             eval_set_file=eval_set_file,
             eval_id=eval_name,
             final_eval_status=final_eval_status,

src/google/adk/cli/cli_tools_click.py

@@ -245,7 +245,7 @@ def cli_eval(
 
   try:
     from .cli_eval import EvalMetric
-    from .cli_eval import EvalResult
+    from .cli_eval import EvalCaseResult
     from .cli_eval import EvalStatus
     from .cli_eval import get_evaluation_criteria_or_default
     from .cli_eval import get_root_agent
@@ -269,7 +269,7 @@ def cli_eval(
 
   eval_set_to_evals = parse_and_get_evals_to_run(eval_set_file_path)
 
-  async def _collect_eval_results() -> list[EvalResult]:
+  async def _collect_eval_results() -> list[EvalCaseResult]:
     return [
         result
         async for result in run_evals(
@@ -290,7 +290,7 @@ def cli_eval(
   eval_run_summary = {}
 
   for eval_result in eval_results:
-    eval_result: EvalResult
+    eval_result: EvalCaseResult
 
     if eval_result.eval_set_file not in eval_run_summary:
       eval_run_summary[eval_result.eval_set_file] = [0, 0]

tests/unittests/cli/utils/test_cli_tools_click.py

@@ -250,7 +250,7 @@ def test_cli_eval_success_path(
     def __init__(self, metric_name: str, threshold: float) -> None:
       ...
 
-  class _EvalResult:
+  class _EvalCaseResult:
 
     def __init__(self, eval_set_file: str, final_eval_status: str) -> None:
       self.eval_set_file = eval_set_file
@@ -261,7 +261,7 @@ def test_cli_eval_success_path(
 
   # helper funcs
   stub.EvalMetric = _EvalMetric
-  stub.EvalResult = _EvalResult
+  stub.EvalCaseResult = _EvalCaseResult
   stub.EvalStatus = _EvalStatus
   stub.MISSING_EVAL_DEPENDENCIES_MESSAGE = "stub msg"
 
@@ -272,8 +272,8 @@ def test_cli_eval_success_path(
 
   # Create an async generator function for run_evals
   async def mock_run_evals(*_a, **_k):
-    yield _EvalResult("set1.json", "PASSED")
+    yield _EvalCaseResult("set1.json", "PASSED")
-    yield _EvalResult("set1.json", "FAILED")
+    yield _EvalCaseResult("set1.json", "FAILED")
 
   stub.run_evals = mock_run_evals