fix: import deprecated from typing_extensions

Fixes https://github.com/google/adk-python/issues/1125

and add it to pyproject.toml. This was added to typing_extensions in 4.5.0. I followed versioning guidance from https://typing-extensions.readthedocs.io/en/latest/index.html#versioning-and-backwards-compatibility.

PiperOrigin-RevId: 767317467

CHANGELOG.md

@@ -1,5 +1,89 @@
 # Changelog
 
+## [1.2.0](https://github.com/google/adk-python/compare/v1.1.1...v1.2.0) (2025-06-04)
+
+
+### Features
+
+* Add agent engine as a deployment option to the ADK CLI ([2409c3e](https://github.com/google/adk-python/commit/2409c3ef192262c80f5328121f6dc4f34265f5cf))
+* Add an option to use gcs artifact service in adk web. ([8d36dbd](https://github.com/google/adk-python/commit/8d36dbda520b1c0dec148e1e1d84e36ddcb9cb95))
+* Add index tracking to handle parallel tool call using litellm ([05f4834](https://github.com/google/adk-python/commit/05f4834759c9b1f0c0af9d89adb7b81ea67d82c8))
+* Add sortByColumn functionality to List Operation ([af95dd2](https://github.com/google/adk-python/commit/af95dd29325865ec30a1945b98e65e457760e003))
+* Add implementation for  `get_eval_case`, `update_eval_case` and `delete_eval_case` for the local eval sets manager. ([a7575e0](https://github.com/google/adk-python/commit/a7575e078a564af6db3f42f650e94ebc4f338918))
+* Expose more config of VertexAiSearchTool from latest Google GenAI SDK ([2b5c89b](https://github.com/google/adk-python/commit/2b5c89b3a94e82ea4a40363ea8de33d9473d7cf0))
+* New Agent Visualization ([da4bc0e](https://github.com/google/adk-python/commit/da4bc0efc0dd96096724559008205854e97c3fd1))
+* Set the max width and height of view image dialog to be 90% ([98a635a](https://github.com/google/adk-python/commit/98a635afee399f64e0a813d681cd8521fbb49500))
+* Support Langchain StructuredTool for Langchain tool ([7e637d3](https://github.com/google/adk-python/commit/7e637d3fa05ca3e43a937e7158008d2b146b1b81))
+* Support Langchain tools that has run_manager in _run args and don't have args_schema populated ([3616bb5](https://github.com/google/adk-python/commit/3616bb5fc4da90e79eb89039fb5e302d6a0a14ec))
+* Update for anthropic models ([16f7d98](https://github.com/google/adk-python/commit/16f7d98acf039f21ec8a99f19eabf0ef4cb5268c))
+* Use bigquery scope by default in bigquery credentials. ([ba5b80d](https://github.com/google/adk-python/commit/ba5b80d5d774ff5fdb61bd43b7849057da2b4edf))
+* Add jira_agent adk samples code which connect Jira cloud ([8759a25](https://github.com/google/adk-python/commit/8759a2525170edb2f4be44236fa646a93ba863e6))
+* Render HTML artifact in chat window ([5c2ad32](https://github.com/google/adk-python/commit/5c2ad327bf4262257c3bc91010c3f8c303d3a5f5))
+* Add export to json button in the chat window ([fc3e374](https://github.com/google/adk-python/commit/fc3e374c86c4de87b4935ee9c56b6259f00e8ea2))
+* Add tooltip to the export session button ([2735942](https://github.com/google/adk-python/commit/273594215efe9dbed44d4ef85e6234bd7ba7b7ae))
+
+
+### Bug Fixes
+
+* Add adk icon for UI ([2623c71](https://github.com/google/adk-python/commit/2623c710868d832b6d5119f38e22d82adb3de66b))
+* Add cache_ok option to remove sa warning. ([841e10a](https://github.com/google/adk-python/commit/841e10ae353e0b1b3d020a26d6cac6f37981550e))
+* Add support for running python main function in UnsafeLocalCodeExecutor when the code has an if __name__ == "__main__" statement. ([95e33ba](https://github.com/google/adk-python/commit/95e33baf57e9c267a758e08108cde76adf8af69b))
+* Adk web not working on some env for windows, fixes https://github.com/google/adk-web/issues/34 ([daac8ce](https://github.com/google/adk-python/commit/daac8cedfe6d894f77ea52784f0a6d19003b2c00))
+* Assign empty inputSchema to MCP tool when converting an ADK tool that wraps a function which takes no parameters. ([2a65c41](https://github.com/google/adk-python/commit/2a65c4118bb2aa97f2a13064db884bd63c14a5f7))
+* Call all tools in parallel calls during partial authentication ([0e72efb](https://github.com/google/adk-python/commit/0e72efb4398ce6a5d782bcdcb770b2473eb5af2e))
+* Continue fetching events if there are multiple pages. ([6506302](https://github.com/google/adk-python/commit/65063023a5a7cb6cd5db43db14a411213dc8acf5))
+* Do not convert "false" value to dict ([60ceea7](https://github.com/google/adk-python/commit/60ceea72bde2143eb102c60cf33b365e1ab07d8f))
+* Enhance agent loader exception handler and expose precise error information ([7b51ae9](https://github.com/google/adk-python/commit/7b51ae97245f6990c089183734aad41fe59b3330))
+* Ensure function description is copied when ignoring parameters ([7fdc6b4](https://github.com/google/adk-python/commit/7fdc6b4417e5cf0fbc72d3117531914353d3984a))
+* Filter memory by app_name and user_id. ([db4bc98](https://github.com/google/adk-python/commit/db4bc9809c7bb6b0d261973ca7cfd87b392694be))
+* Fix filtering by user_id for vertex ai session service listing ([9d4ca4e](https://github.com/google/adk-python/commit/9d4ca4ed44cf10bc87f577873faa49af469acc25))
+* fix parameter schema generation for gemini ([5a67a94](https://github.com/google/adk-python/commit/5a67a946d2168b80dd6eba008218468c2db2e74e))
+* Handle non-indexed function call chunks with incremental fallback index ([b181cbc](https://github.com/google/adk-python/commit/b181cbc8bc629d1c9bfd50054e47a0a1b04f7410))
+* Handles function tool parsing corner case where type hints are stored as strings. ([a8a2074](https://github.com/google/adk-python/commit/a8a20743f92cd63c3d287a3d503c1913dd5ad5ae))
+* Introduce PreciseTimestamp to fix mysql datetime precision issue. ([841e10a](https://github.com/google/adk-python/commit/841e10ae353e0b1b3d020a26d6cac6f37981550e))
+* match arg case in errors ([b226a06](https://github.com/google/adk-python/commit/b226a06c0bf798f85a53c591ad12ee582703af6d))
+* ParallelAgent should only append to its immediate sub-agent, not transitive descendants ([ec8bc73](https://github.com/google/adk-python/commit/ec8bc7387c84c3f261c44cedfe76eb1f702e7b17))
+* Relax openapi spec to gemini schema conversion to tolerate more cases ([b1a74d0](https://github.com/google/adk-python/commit/b1a74d099fae44d41750b79e58455282d919dd78))
+* Remove labels from config when using API key from Google AI Studio to call model ([5d29716](https://github.com/google/adk-python/commit/5d297169d08a2d0ea1a07641da2ac39fa46b68a4))
+* **sample:** Correct text artifact saving in artifact_save_text sample ([5c6001d](https://github.com/google/adk-python/commit/5c6001d90fe6e1d15a2db6b30ecf9e7b6c26eee4))
+* Separate thinking from text parts in streaming mode ([795605a](https://github.com/google/adk-python/commit/795605a37e1141e37d86c9b3fa484a3a03e7e9a6))
+* Simplify content for ollama provider ([eaee49b](https://github.com/google/adk-python/commit/eaee49bc897c20231ecacde6855cccfa5e80d849))
+* Timeout issues for mcpstdio server when mcp tools are incorrect. ([45ef668](https://github.com/google/adk-python/commit/45ef6684352e3c8082958bece8610df60048f4a3))
+* **transfer_to_agent:** update docstring for clarity and accuracy ([854a544](https://github.com/google/adk-python/commit/854a5440614590c2a3466cf652688ba57d637205))
+* Update unit test code for test_connection ([b0403b2](https://github.com/google/adk-python/commit/b0403b2d98b2776d15475f6b525409670e2841fc))
+* Use inspect.cleandoc on function docstrings in generate_function_declaration. ([f7cb666](https://github.com/google/adk-python/commit/f7cb66620be843b8d9f3d197d6e8988e9ee0dfca))
+* Restore errors path ([32c5ffa](https://github.com/google/adk-python/commit/32c5ffa8ca5e037f41ff345f9eecf5b26f926ea1))
+* Unused import for deprecated ([ccd05e0](https://github.com/google/adk-python/commit/ccd05e0b00d0327186e3b1156f1b0216293efe21))
+* Prevent JSON parsing errors and preserve non-ascii characters in telemetry ([d587270](https://github.com/google/adk-python/commit/d587270327a8de9f33b3268de5811ac756959850))
+* Raise HTTPException when running evals in fast_api if google-adk[eval] is not installed ([1de5c34](https://github.com/google/adk-python/commit/1de5c340d8da1cedee223f6f5a8c90070a9f0298))
+* Fix typos in README for sample bigquery_agent and oauth_calendar_agent ([9bdd813](https://github.com/google/adk-python/commit/9bdd813be15935af5c5d2a6982a2391a640cab23))
+* Make tool_call one span for telemetry and renamed to execute_tool ([999a7fe](https://github.com/google/adk-python/commit/999a7fe69d511b1401b295d23ab3c2f40bccdc6f))
+* Use media type in chat window. Remove isArtifactImage and isArtifactAudio reference ([1452dac](https://github.com/google/adk-python/commit/1452dacfeb6b9970284e1ddeee6c4f3cb56781f8))
+* Set output_schema correctly for LiteLllm ([6157db7](https://github.com/google/adk-python/commit/6157db77f2fba4a44d075b51c83bff844027a147))
+* Update pending event dialog style ([1db601c](https://github.com/google/adk-python/commit/1db601c4bd90467b97a2f26fe9d90d665eb3c740))
+* Remove the gap between event holder and image ([63822c3](https://github.com/google/adk-python/commit/63822c3fa8b0bdce2527bd0d909c038e2b66dd98))
+
+
+### Documentation
+
+* Adds a sample agent to illustrate state usage via `callbacks`. ([18fbe3c](https://github.com/google/adk-python/commit/18fbe3cbfc9f2af97e4b744ec0a7552331b1d8e3))
+* Fix typos in documentation ([7aaf811](https://github.com/google/adk-python/commit/7aaf8116169c210ceda35c649b5b49fb65bbb740))
+* Change eval_dataset to eval_dataset_file_path_or_dir ([62d7bf5](https://github.com/google/adk-python/commit/62d7bf58bb1c874caaf3c56a614500ae3b52f215))
+* Fix broken link to A2A example ([0d66a78](https://github.com/google/adk-python/commit/0d66a7888b68380241b92f7de394a06df5a0cc06))
+* Fix typo in envs.py ([bd588bc](https://github.com/google/adk-python/commit/bd588bce50ccd0e70b96c7291db035a327ad4d24))
+* Updates CONTRIBUTING.md to refine setup process using uv. ([04e07b4](https://github.com/google/adk-python/commit/04e07b4a1451123272641a256c6af1528ea6523e))
+* Create and update project documentation including README.md and CONTRIBUTING.md ([f180331](https://github.com/google/adk-python/commit/f1803312c6a046f94c23cfeaed3e8656afccf7c3))
+* Rename the root agent in the example to match the example name ([94c0aca](https://github.com/google/adk-python/commit/94c0aca685f1dfa4edb44caaedc2de25cc0caa41))
+* ADK: add section comment ([349a414](https://github.com/google/adk-python/commit/349a414120fbff0937966af95864bd683f063d08))
+
+
+### Chore
+
+* Miscellaneous changes ([0724a83](https://github.com/google/adk-python/commit/0724a83aa9cda00c1b228ed47a5baa7527bb4a0a), [a9dcc58](https://github.com/google/adk-python/commit/a9dcc588ad63013d063dbe37095c0d2e870142c3), [ac52eab](https://github.com/google/adk-python/commit/ac52eab88eccafa451be7584e24aea93ff15f3f3), [a0714b8](https://github.com/google/adk-python/commit/a0714b8afc55461f315ede8451b17aad18d698dd))
+* Enable release-please workflow ([57d99aa](https://github.com/google/adk-python/commit/57d99aa7897fb229f41c2a08034606df1e1e6064))
+* Added unit test coverage for local_eval_sets_manager.py ([174afb3](https://github.com/google/adk-python/commit/174afb3975bdc7e5f10c26f3eebb17d2efa0dd59))
+* Extract common options for `adk web` and `adk api_server` ([01965bd](https://github.com/google/adk-python/commit/01965bdd74a9dbdb0ce91a924db8dee5961478b8))
+
 ## 1.1.1
 
 ### Features

contributing/samples/hello_world_anthropic/agent.py

@@ -61,7 +61,7 @@ async def check_prime(nums: list[int]) -> str:
 
 root_agent = Agent(
     model=Claude(model="claude-3-5-sonnet-v2@20241022"),
-    name="data_processing_agent",
+    name="hello_world_agent",
     description=(
         "hello world agent that can roll a dice of 8 sides and check prime"
         " numbers."

pyproject.toml

@@ -44,6 +44,7 @@ dependencies = [
   "PyYAML>=6.0.2",                                   # For APIHubToolset.
   "sqlalchemy>=2.0",                                 # SQL database ORM
   "tzlocal>=5.3",                                    # Time zone utilities
+  "typing-extensions>=4.5, <5",
   "uvicorn>=0.34.0",                                 # ASGI server for FastAPI
   # go/keep-sorted end
 ]

src/google/adk/cli/utils/evals.py

@@ -12,18 +12,20 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
+from __future__ import annotations
+
 from typing import Any
 from typing import Tuple
 
-from deprecated import deprecated
 from google.genai import types as genai_types
+from typing_extensions import deprecated
 
 from ...evaluation.eval_case import IntermediateData
 from ...evaluation.eval_case import Invocation
 from ...sessions.session import Session
 
 
-@deprecated(reason='Use convert_session_to_eval_invocations instead.')
+@deprecated('Use convert_session_to_eval_invocations instead.')
 def convert_session_to_eval_format(session: Session) -> list[dict[str, Any]]:
   """Converts a session data into eval format.
 

src/google/adk/evaluation/response_evaluator.py

@@ -12,13 +12,15 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
+from __future__ import annotations
+
 from typing import Any
 from typing import Optional
 
-from deprecated import deprecated
 from google.genai import types as genai_types
 import pandas as pd
 from tabulate import tabulate
+from typing_extensions import deprecated
 from typing_extensions import override
 from vertexai.preview.evaluation import EvalTask
 from vertexai.preview.evaluation import MetricPromptTemplateExamples
@@ -124,10 +126,8 @@ class ResponseEvaluator(Evaluator):
 
   @staticmethod
   @deprecated(
-      reason=(
-          "This method has been deprecated and will be removed soon. Please use"
-          " evaluate_invocations instead."
-      )
+      "This method has been deprecated and will be removed soon. Please use"
+      " evaluate_invocations instead."
   )
   def evaluate(
       raw_eval_dataset: list[list[dict[str, Any]]],

src/google/adk/evaluation/trajectory_evaluator.py

@@ -12,13 +12,15 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
+from __future__ import annotations
+
 from typing import Any
 from typing import cast
 
-from deprecated import deprecated
 from google.genai import types as genai_types
 import pandas as pd
 from tabulate import tabulate
+from typing_extensions import deprecated
 from typing_extensions import override
 
 from .eval_case import Invocation
@@ -100,10 +102,8 @@ class TrajectoryEvaluator(Evaluator):
 
   @staticmethod
   @deprecated(
-      reason=(
-          "This method has been deprecated and will be removed soon. Please use"
-          " evaluate_invocations instead."
-      )
+      "This method has been deprecated and will be removed soon. Please use"
+      " evaluate_invocations instead."
   )
   def evaluate(
       eval_dataset: list[list[dict[str, Any]]],
@@ -218,7 +218,10 @@ class TrajectoryEvaluator(Evaluator):
     return new_row, failure
 
   @staticmethod
-  @deprecated()
+  @deprecated(
+      "are_tools_equal is deprecated and will be removed soon. Please use"
+      " TrajectoryEvaluator._are_tool_calls_equal instead."
+  )
   def are_tools_equal(list_a_original, list_b_original):
     # Remove other entries that we don't want to evaluate
     list_a = [

src/google/adk/sessions/in_memory_session_service.py

@@ -11,6 +11,7 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
+from __future__ import annotations
 
 import copy
 import logging
@@ -223,6 +224,7 @@ class InMemorySessionService(BaseSessionService):
       sessions_without_events.append(copied_session)
     return ListSessionsResponse(sessions=sessions_without_events)
 
+  @override
   async def delete_session(
       self, *, app_name: str, user_id: str, session_id: str
   ) -> None:
@@ -247,7 +249,7 @@ class InMemorySessionService(BaseSessionService):
         )
         is None
     ):
-      return None
+      return
 
     self.sessions[app_name][user_id].pop(session_id)
 
@@ -261,11 +263,20 @@ class InMemorySessionService(BaseSessionService):
     app_name = session.app_name
     user_id = session.user_id
     session_id = session.id
+
+    def _warning(message: str) -> None:
+      logger.warning(
+          f'Failed to append event to session {session_id}: {message}'
+      )
+
     if app_name not in self.sessions:
+      _warning(f'app_name {app_name} not in sessions')
       return event
     if user_id not in self.sessions[app_name]:
+      _warning(f'user_id {user_id} not in sessions[app_name]')
       return event
     if session_id not in self.sessions[app_name][user_id]:
+      _warning(f'session_id {session_id} not in sessions[app_name][user_id]')
       return event
 
     if event.actions and event.actions.state_delta:

src/google/adk/sessions/vertex_ai_session_service.py

@@ -16,8 +16,8 @@ from __future__ import annotations
 import asyncio
 import logging
 import re
-import time
 from typing import Any
+from typing import Dict
 from typing import Optional
 import urllib.parse
 
@@ -50,9 +50,6 @@ class VertexAiSessionService(BaseSessionService):
     self.project = project
     self.location = location
 
-    client = genai.Client(vertexai=True, project=project, location=location)
-    self.api_client = client._api_client
-
   @override
   async def create_session(
       self,
@@ -86,6 +83,7 @@ class VertexAiSessionService(BaseSessionService):
     operation_id = api_response['name'].split('/')[-1]
 
     max_retry_attempt = 5
+    lro_response = None
     while max_retry_attempt >= 0:
       lro_response = await api_client.async_request(
           http_method='GET',
@@ -99,6 +97,11 @@ class VertexAiSessionService(BaseSessionService):
       await asyncio.sleep(1)
       max_retry_attempt -= 1
 
+    if lro_response is None or not lro_response.get('done', None):
+      raise TimeoutError(
+          f'Timeout waiting for operation {operation_id} to complete.'
+      )
+
     # Get session resource
     get_session_api_response = await api_client.async_request(
         http_method='GET',
@@ -235,11 +238,15 @@ class VertexAiSessionService(BaseSessionService):
   ) -> None:
     reasoning_engine_id = _parse_reasoning_engine_id(app_name)
     api_client = _get_api_client(self.project, self.location)
-    await api_client.async_request(
-        http_method='DELETE',
-        path=f'reasoningEngines/{reasoning_engine_id}/sessions/{session_id}',
-        request_dict={},
-    )
+    try:
+      await api_client.async_request(
+          http_method='DELETE',
+          path=f'reasoningEngines/{reasoning_engine_id}/sessions/{session_id}',
+          request_dict={},
+      )
+    except Exception as e:
+      logger.error(f'Error deleting session {session_id}: {e}')
+      raise e
 
   @override
   async def append_event(self, session: Session, event: Event) -> Event:
@@ -266,7 +273,7 @@ def _get_api_client(project: str, location: str):
   return client._api_client
 
 
-def _convert_event_to_json(event: Event):
+def _convert_event_to_json(event: Event) -> Dict[str, Any]:
   metadata_json = {
       'partial': event.partial,
       'turn_complete': event.turn_complete,
@@ -318,7 +325,7 @@ def _convert_event_to_json(event: Event):
   return event_json
 
 
-def _from_api_event(api_event: dict) -> Event:
+def _from_api_event(api_event: Dict[str, Any]) -> Event:
   event_actions = EventActions()
   if api_event.get('actions', None):
     event_actions = EventActions(

src/google/adk/tools/_forwarding_artifact_service.py

@@ -0,0 +1,96 @@
+# Copyright 2025 Google LLC
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+from __future__ import annotations
+
+from typing import Optional
+from typing import TYPE_CHECKING
+
+from google.genai import types
+from typing_extensions import override
+
+from ..artifacts.base_artifact_service import BaseArtifactService
+
+if TYPE_CHECKING:
+  from .tool_context import ToolContext
+
+
+class ForwardingArtifactService(BaseArtifactService):
+  """Artifact service that forwards to the parent tool context."""
+
+  def __init__(self, tool_context: ToolContext):
+    self.tool_context = tool_context
+    self._invocation_context = tool_context._invocation_context
+
+  @override
+  async def save_artifact(
+      self,
+      *,
+      app_name: str,
+      user_id: str,
+      session_id: str,
+      filename: str,
+      artifact: types.Part,
+  ) -> int:
+    return await self.tool_context.save_artifact(
+        filename=filename, artifact=artifact
+    )
+
+  @override
+  async def load_artifact(
+      self,
+      *,
+      app_name: str,
+      user_id: str,
+      session_id: str,
+      filename: str,
+      version: Optional[int] = None,
+  ) -> Optional[types.Part]:
+    return await self.tool_context.load_artifact(
+        filename=filename, version=version
+    )
+
+  @override
+  async def list_artifact_keys(
+      self, *, app_name: str, user_id: str, session_id: str
+  ) -> list[str]:
+    return await self.tool_context.list_artifacts()
+
+  @override
+  async def delete_artifact(
+      self, *, app_name: str, user_id: str, session_id: str, filename: str
+  ) -> None:
+    del app_name, user_id, session_id
+    if self._invocation_context.artifact_service is None:
+      raise ValueError("Artifact service is not initialized.")
+    await self._invocation_context.artifact_service.delete_artifact(
+        app_name=self._invocation_context.app_name,
+        user_id=self._invocation_context.user_id,
+        session_id=self._invocation_context.session.id,
+        filename=filename,
+    )
+
+  @override
+  async def list_versions(
+      self, *, app_name: str, user_id: str, session_id: str, filename: str
+  ) -> list[int]:
+    del app_name, user_id, session_id
+    if self._invocation_context.artifact_service is None:
+      raise ValueError("Artifact service is not initialized.")
+    return await self._invocation_context.artifact_service.list_versions(
+        app_name=self._invocation_context.app_name,
+        user_id=self._invocation_context.user_id,
+        session_id=self._invocation_context.session.id,
+        filename=filename,
+    )

src/google/adk/tools/_function_parameter_parse_util.py

@@ -37,6 +37,7 @@ _py_builtin_type_to_schema_type = {
     bool: types.Type.BOOLEAN,
     list: types.Type.ARRAY,
     dict: types.Type.OBJECT,
+    None: types.Type.NULL,
 }
 
 logger = logging.getLogger('google_adk.' + __name__)

src/google/adk/tools/agent_tool.py

@@ -25,6 +25,7 @@ from . import _automatic_function_calling_util
 from ..memory.in_memory_memory_service import InMemoryMemoryService
 from ..runners import Runner
 from ..sessions.in_memory_session_service import InMemorySessionService
+from ._forwarding_artifact_service import ForwardingArtifactService
 from .base_tool import BaseTool
 from .tool_context import ToolContext
 
@@ -123,9 +124,7 @@ class AgentTool(BaseTool):
     runner = Runner(
         app_name=self.agent.name,
         agent=self.agent,
-        # TODO(kech): Remove the access to the invocation context.
-        #   It seems we don't need re-use artifact_service if we forward below.
-        artifact_service=tool_context._invocation_context.artifact_service,
+        artifact_service=ForwardingArtifactService(tool_context),
         session_service=InMemorySessionService(),
         memory_service=InMemoryMemoryService(),
     )
@@ -144,24 +143,6 @@ class AgentTool(BaseTool):
         tool_context.state.update(event.actions.state_delta)
       last_event = event
 
-    if runner.artifact_service:
-      # Forward all artifacts to parent session.
-      artifact_names = await runner.artifact_service.list_artifact_keys(
-          app_name=session.app_name,
-          user_id=session.user_id,
-          session_id=session.id,
-      )
-      for artifact_name in artifact_names:
-        if artifact := await runner.artifact_service.load_artifact(
-            app_name=session.app_name,
-            user_id=session.user_id,
-            session_id=session.id,
-            filename=artifact_name,
-        ):
-          await tool_context.save_artifact(
-              filename=artifact_name, artifact=artifact
-          )
-
     if not last_event or not last_event.content or not last_event.content.parts:
       return ''
     if isinstance(self.agent, LlmAgent) and self.agent.output_schema:

src/google/adk/version.py

@@ -12,5 +12,5 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-# version: date+base_cl
-__version__ = "1.1.1"
+# version: major.minor.patch
+__version__ = "1.2.0"

tests/unittests/evaluation/test_local_eval_set_results_manager.py

@@ -0,0 +1,168 @@
+# Copyright 2025 Google LLC
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+from __future__ import annotations
+
+import json
+import os
+import shutil
+import tempfile
+import time
+from unittest.mock import patch
+
+from google.adk.evaluation.eval_result import EvalCaseResult
+from google.adk.evaluation.eval_result import EvalSetResult
+from google.adk.evaluation.evaluator import EvalStatus
+from google.adk.evaluation.local_eval_set_results_manager import _ADK_EVAL_HISTORY_DIR
+from google.adk.evaluation.local_eval_set_results_manager import _EVAL_SET_RESULT_FILE_EXTENSION
+from google.adk.evaluation.local_eval_set_results_manager import _sanitize_eval_set_result_name
+from google.adk.evaluation.local_eval_set_results_manager import LocalEvalSetResultsManager
+import pytest
+
+
+def test_sanitize_eval_set_result_name():
+  assert _sanitize_eval_set_result_name("app/name") == "app_name"
+  assert _sanitize_eval_set_result_name("app_name") == "app_name"
+  assert _sanitize_eval_set_result_name("app/name/with/slashes") == (
+      "app_name_with_slashes"
+  )
+
+
+class TestLocalEvalSetResultsManager:
+
+  @pytest.fixture(autouse=True)
+  def setup(self):
+    self.temp_dir = tempfile.mkdtemp()
+    self.agents_dir = os.path.join(self.temp_dir, "agents")
+    os.makedirs(self.agents_dir)
+    self.manager = LocalEvalSetResultsManager(self.agents_dir)
+    self.app_name = "test_app"
+    self.eval_set_id = "test_eval_set"
+    self.eval_case_results = [
+        EvalCaseResult(
+            eval_set_file="test_file",
+            eval_set_id=self.eval_set_id,
+            eval_id="case1",
+            final_eval_status=EvalStatus.PASSED,
+            eval_metric_results=[],
+            overall_eval_metric_results=[],
+            eval_metric_result_per_invocation=[],
+            session_id="session1",
+        )
+    ]
+    self.timestamp = time.time()  # Store the timestamp
+    self.eval_set_result_id = (
+        self.app_name + "_" + self.eval_set_id + "_" + str(self.timestamp)
+    )
+    self.eval_set_result_name = _sanitize_eval_set_result_name(
+        self.eval_set_result_id
+    )
+    self.eval_set_result = EvalSetResult(
+        eval_set_result_id=self.eval_set_result_id,
+        eval_set_result_name=self.eval_set_result_name,
+        eval_set_id=self.eval_set_id,
+        eval_case_results=self.eval_case_results,
+        creation_timestamp=self.timestamp,
+    )
+
+  def teardown(self):
+    shutil.rmtree(self.temp_dir)
+
+  @patch("time.time")
+  def test_save_eval_set_result(self, mock_time):
+    mock_time.return_value = self.timestamp
+    self.manager.save_eval_set_result(
+        self.app_name, self.eval_set_id, self.eval_case_results
+    )
+    eval_history_dir = os.path.join(
+        self.agents_dir, self.app_name, _ADK_EVAL_HISTORY_DIR
+    )
+    expected_file_path = os.path.join(
+        eval_history_dir,
+        self.eval_set_result_name + _EVAL_SET_RESULT_FILE_EXTENSION,
+    )
+    assert os.path.exists(expected_file_path)
+    with open(expected_file_path, "r") as f:
+      actual_eval_set_result_json = json.load(f)
+
+    # need to convert eval_set_result to json
+    expected_eval_set_result_json = self.eval_set_result.model_dump_json()
+    assert expected_eval_set_result_json == actual_eval_set_result_json
+
+  @patch("time.time")
+  def test_get_eval_set_result(self, mock_time):
+    mock_time.return_value = self.timestamp
+    self.manager.save_eval_set_result(
+        self.app_name, self.eval_set_id, self.eval_case_results
+    )
+    retrieved_result = self.manager.get_eval_set_result(
+        self.app_name, self.eval_set_result_name
+    )
+    assert retrieved_result == self.eval_set_result
+
+  @patch("time.time")
+  def test_get_eval_set_result_not_found(self, mock_time):
+    mock_time.return_value = self.timestamp
+
+    with pytest.raises(ValueError) as e:
+      self.manager.get_eval_set_result(self.app_name, "non_existent_id")
+
+    assert "does not exist" in str(e.value)
+
+  @patch("time.time")
+  def test_list_eval_set_results(self, mock_time):
+    mock_time.return_value = self.timestamp
+    # Save two eval set results for the same app
+    self.manager.save_eval_set_result(
+        self.app_name, self.eval_set_id, self.eval_case_results
+    )
+    timestamp2 = time.time() + 1
+    mock_time.return_value = timestamp2
+    eval_set_result_id2 = (
+        self.app_name + "_" + self.eval_set_id + "_" + str(timestamp2)
+    )
+    eval_set_result_name2 = _sanitize_eval_set_result_name(eval_set_result_id2)
+    eval_case_results2 = [
+        EvalCaseResult(
+            eval_set_file="test_file",
+            eval_set_id=self.eval_set_id,
+            eval_id="case2",
+            final_eval_status=EvalStatus.FAILED,
+            eval_metric_results=[],
+            overall_eval_metric_results=[],
+            eval_metric_result_per_invocation=[],
+            session_id="session2",
+        )
+    ]
+    self.manager.save_eval_set_result(
+        self.app_name, self.eval_set_id, eval_case_results2
+    )
+
+    # Save one eval set result for a different app
+    app_name2 = "another_app"
+    timestamp3 = time.time() + 2
+    mock_time.return_value = timestamp3
+
+    self.manager.save_eval_set_result(
+        app_name2, self.eval_set_id, self.eval_case_results
+    )
+
+    results = self.manager.list_eval_set_results(self.app_name)
+    expected_result = [self.eval_set_result_name, eval_set_result_name2]
+    assert set(results) == set(expected_result)
+
+  def test_list_eval_set_results_empty(self):
+    # No eval set results saved for the app
+    results = self.manager.list_eval_set_results(self.app_name)
+    assert results == []

tests/unittests/tools/test_agent_tool.py

@@ -13,20 +13,15 @@
 # limitations under the License.
 
 from google.adk.agents import Agent
+from google.adk.agents import SequentialAgent
 from google.adk.agents.callback_context import CallbackContext
 from google.adk.tools.agent_tool import AgentTool
 from google.genai.types import Part
 from pydantic import BaseModel
-import pytest
 from pytest import mark
 
 from .. import testing_utils
 
-pytestmark = pytest.mark.skip(
-    reason='Skipping until tool.func evaluations are fixed (async)'
-)
-
-
 function_call_custom = Part.from_function_call(
     name='tool_agent', args={'custom_input': 'test1'}
 )
@@ -112,6 +107,55 @@ def test_update_state():
   assert runner.session.state['state_1'] == 'changed_value'
 
 
+def test_update_artifacts():
+  """The agent tool can read and write artifacts."""
+
+  async def before_tool_agent(callback_context: CallbackContext):
+    # Artifact 1 should be available in the tool agent.
+    artifact = await callback_context.load_artifact('artifact_1')
+    await callback_context.save_artifact(
+        'artifact_2', Part.from_text(text=artifact.text + ' 2')
+    )
+
+  tool_agent = SequentialAgent(
+      name='tool_agent',
+      before_agent_callback=before_tool_agent,
+  )
+
+  async def before_main_agent(callback_context: CallbackContext):
+    await callback_context.save_artifact(
+        'artifact_1', Part.from_text(text='test')
+    )
+
+  async def after_main_agent(callback_context: CallbackContext):
+    # Artifact 2 should be available after the tool agent.
+    artifact_2 = await callback_context.load_artifact('artifact_2')
+    await callback_context.save_artifact(
+        'artifact_3', Part.from_text(text=artifact_2.text + ' 3')
+    )
+
+  mock_model = testing_utils.MockModel.create(
+      responses=[function_call_no_schema, 'response2']
+  )
+  root_agent = Agent(
+      name='root_agent',
+      before_agent_callback=before_main_agent,
+      after_agent_callback=after_main_agent,
+      tools=[AgentTool(agent=tool_agent)],
+      model=mock_model,
+  )
+
+  runner = testing_utils.InMemoryRunner(root_agent)
+  runner.run('test1')
+
+  artifacts_path = f'test_app/test_user/{runner.session_id}'
+  assert runner.runner.artifact_service.artifacts == {
+      f'{artifacts_path}/artifact_1': [Part.from_text(text='test')],
+      f'{artifacts_path}/artifact_2': [Part.from_text(text='test 2')],
+      f'{artifacts_path}/artifact_3': [Part.from_text(text='test 2 3')],
+  }
+
+
 @mark.parametrize(
     'env_variables',
     [