chore: Enhance a2a part converters

a. fix binary data conversion
b. support thoughts, code execution result, executable codes conversion

PiperOrigin-RevId: 775827259

.github/ISSUE_TEMPLATE/bug_report.md

@@ -31,5 +31,8 @@ If applicable, add screenshots to help explain your problem.
  - Python version(python -V):
  - ADK version(pip show google-adk):
 
+ **Model Information:**
+ For example, which model is being used.
+
 **Additional context**
 Add any other context about the problem here.

.github/workflows/python-unit-tests.yml

@@ -36,8 +36,8 @@ jobs:
       with:
         python-version: ${{ matrix.python-version }}
 
-    - name: Install uv
-      run: curl -LsSf https://astral.sh/uv/install.sh | sh
+    - name: Install the latest version of uv
+      uses: astral-sh/setup-uv@v6
 
     - name: Install dependencies
       run: |

.gitignore

@@ -82,6 +82,7 @@ log/
 .env.development.local
 .env.test.local
 .env.production.local
+uv.lock
 
 # Google Cloud specific
 .gcloudignore

CHANGELOG.md

@@ -1,5 +1,40 @@
 # Changelog
 
+## [1.5.0](https://github.com/google/adk-python/compare/v1.4.2...v1.5.0) (2025-06-25)
+
+
+### Features
+
+* Add a new option `eval_storage_uri` in adk web & adk eval to specify GCS bucket to store eval data ([fa025d7](https://github.com/google/adk-python/commit/fa025d755978e1506fa0da1fecc49775bebc1045))
+* Add ADK examples for litellm with add_function_to_prompt ([f33e090](https://github.com/google/adk-python/commit/f33e0903b21b752168db3006dd034d7d43f7e84d))
+* Add implementation of VertexAiMemoryBankService and support in FastAPI endpoint ([abc89d2](https://github.com/google/adk-python/commit/abc89d2c811ba00805f81b27a3a07d56bdf55a0b))
+* Add rouge_score library to ADK eval dependencies, and implement RougeEvaluator that is computes ROUGE-1 for "response_match_score" metric ([9597a44](https://github.com/google/adk-python/commit/9597a446fdec63ad9e4c2692d6966b14f80ff8e2))
+* Add usage span attributes to telemetry ([#356](https://github.com/google/adk-python/issues/356)) ([ea69c90](https://github.com/google/adk-python/commit/ea69c9093a16489afdf72657136c96f61c69cafd))
+* Add Vertex Express mode compatibility for VertexAiSessionService ([00cc8cd](https://github.com/google/adk-python/commit/00cc8cd6433fc45ecfc2dbaa04dbbc1a81213b4d))
+
+
+### Bug Fixes
+
+* Include current turn context when include_contents='none' ([9e473e0](https://github.com/google/adk-python/commit/9e473e0abdded24e710fd857782356c15d04b515))
+* Make LiteLLM streaming truly asynchronous ([bd67e84](https://github.com/google/adk-python/commit/bd67e8480f6e8b4b0f8c22b94f15a8cda1336339))
+* Make raw_auth_credential and exchanged_auth_credential optional given their default value is None ([acbdca0](https://github.com/google/adk-python/commit/acbdca0d8400e292ba5525931175e0d6feab15f1))
+* Minor typo fix in the agent instruction ([ef3c745](https://github.com/google/adk-python/commit/ef3c745d655538ebd1ed735671be615f842341a8))
+* Typo fix in sample agent instruction ([ef3c745](https://github.com/google/adk-python/commit/ef3c745d655538ebd1ed735671be615f842341a8))
+* Update contributing links ([a1e1441](https://github.com/google/adk-python/commit/a1e14411159fd9f3e114e15b39b4949d0fd6ecb1))
+* Use starred tuple unpacking on GCS artifact blob names ([3b1d9a8](https://github.com/google/adk-python/commit/3b1d9a8a3e631ca2d86d30f09640497f1728986c))
+
+
+### Chore
+
+* Do not send api request when session does not have events ([88a4402](https://github.com/google/adk-python/commit/88a4402d142672171d0a8ceae74671f47fa14289))
+* Leverage official uv action for install([09f1269](https://github.com/google/adk-python/commit/09f1269bf7fa46ab4b9324e7f92b4f70ffc923e5))
+* Update google-genai package and related deps to latest([ed7a21e](https://github.com/google/adk-python/commit/ed7a21e1890466fcdf04f7025775305dc71f603d))
+* Add credential service backed by session state([29cd183](https://github.com/google/adk-python/commit/29cd183aa1b47dc4f5d8afe22f410f8546634abc))
+* Clarify the behavior of Event.invocation_id([f033e40](https://github.com/google/adk-python/commit/f033e405c10ff8d86550d1419a9d63c0099182f9))
+* Send user message to the agent that returned a corresponding function call if user message is a function response([7c670f6](https://github.com/google/adk-python/commit/7c670f638bc17374ceb08740bdd057e55c9c2e12))
+* Add request converter to convert a2a request to ADK request([fb13963](https://github.com/google/adk-python/commit/fb13963deda0ff0650ac27771711ea0411474bf5))
+* Support allow_origins in cloud_run deployment ([2fd8feb](https://github.com/google/adk-python/commit/2fd8feb65d6ae59732fb3ec0652d5650f47132cc))
+
 ## [1.4.2](https://github.com/google/adk-python/compare/v1.4.1...v1.4.2) (2025-06-20)
 
 

CONTRIBUTING.md

@@ -200,7 +200,7 @@ For any changes that impact user-facing documentation (guides, API reference, tu
 
 ## Contributing Resources
 
-[Contributing folder](https://github.com/google/adk-python/tree/main/contributing/samples) has resources that is helpful for contributors.
+[Contributing folder](https://github.com/google/adk-python/tree/main/contributing) has resources that is helpful for contributors.
 
 
 ## Code reviews

README.md

@@ -135,7 +135,7 @@ adk eval \
 ## 🤝 Contributing
 
 We welcome contributions from the community! Whether it's bug reports, feature requests, documentation improvements, or code contributions, please see our
-- [General contribution guideline and flow](https://google.github.io/adk-docs/contributing-guide/#questions).
+- [General contribution guideline and flow](https://google.github.io/adk-docs/contributing-guide/).
 - Then if you want to contribute code, please read [Code Contributing Guidelines](./CONTRIBUTING.md) to get started.
 
 ## 📄 License

contributing/samples/adk_issue_formatting_agent/agent.py

@@ -0,0 +1,241 @@
+# 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 pathlib import Path
+from typing import Any
+
+from adk_issue_formatting_agent.settings import GITHUB_BASE_URL
+from adk_issue_formatting_agent.settings import IS_INTERACTIVE
+from adk_issue_formatting_agent.settings import OWNER
+from adk_issue_formatting_agent.settings import REPO
+from adk_issue_formatting_agent.utils import error_response
+from adk_issue_formatting_agent.utils import get_request
+from adk_issue_formatting_agent.utils import post_request
+from adk_issue_formatting_agent.utils import read_file
+from google.adk import Agent
+import requests
+
+BUG_REPORT_TEMPLATE = read_file(
+    Path(__file__).parent / "../../../../.github/ISSUE_TEMPLATE/bug_report.md"
+)
+FREATURE_REQUEST_TEMPLATE = read_file(
+    Path(__file__).parent
+    / "../../../../.github/ISSUE_TEMPLATE/feature_request.md"
+)
+
+APPROVAL_INSTRUCTION = (
+    "**Do not** wait or ask for user approval or confirmation for adding the"
+    " comment."
+)
+if IS_INTERACTIVE:
+  APPROVAL_INSTRUCTION = (
+      "Ask for user approval or confirmation for adding the comment."
+  )
+
+
+def list_open_issues(issue_count: int) -> dict[str, Any]:
+  """List most recent `issue_count` numer of open issues in the repo.
+
+  Args:
+    issue_count: number of issues to return
+
+  Returns:
+    The status of this request, with a list of issues when successful.
+  """
+  url = f"{GITHUB_BASE_URL}/search/issues"
+  query = f"repo:{OWNER}/{REPO} is:open is:issue"
+  params = {
+      "q": query,
+      "sort": "created",
+      "order": "desc",
+      "per_page": issue_count,
+      "page": 1,
+  }
+
+  try:
+    response = get_request(url, params)
+  except requests.exceptions.RequestException as e:
+    return error_response(f"Error: {e}")
+  issues = response.get("items", None)
+  return {"status": "success", "issues": issues}
+
+
+def get_issue(issue_number: int) -> dict[str, Any]:
+  """Get the details of the specified issue number.
+
+  Args:
+    issue_number: issue number of the Github issue.
+
+  Returns:
+    The status of this request, with the issue details when successful.
+  """
+  url = f"{GITHUB_BASE_URL}/repos/{OWNER}/{REPO}/issues/{issue_number}"
+  try:
+    response = get_request(url)
+  except requests.exceptions.RequestException as e:
+    return error_response(f"Error: {e}")
+  return {"status": "success", "issue": response}
+
+
+def add_comment_to_issue(issue_number: int, comment: str) -> dict[str, any]:
+  """Add the specified comment to the given issue number.
+
+  Args:
+    issue_number: issue number of the Github issue
+    comment: comment to add
+
+  Returns:
+    The the status of this request, with the applied comment when successful.
+  """
+  print(f"Attempting to add comment '{comment}' to issue #{issue_number}")
+  url = f"{GITHUB_BASE_URL}/repos/{OWNER}/{REPO}/issues/{issue_number}/comments"
+  payload = {"body": comment}
+
+  try:
+    response = post_request(url, payload)
+  except requests.exceptions.RequestException as e:
+    return error_response(f"Error: {e}")
+  return {
+      "status": "success",
+      "added_comment": response,
+  }
+
+
+def list_comments_on_issue(issue_number: int) -> dict[str, any]:
+  """List all comments on the given issue number.
+
+  Args:
+    issue_number: issue number of the Github issue
+
+  Returns:
+    The the status of this request, with the list of comments when successful.
+  """
+  print(f"Attempting to list comments on issue #{issue_number}")
+  url = f"{GITHUB_BASE_URL}/repos/{OWNER}/{REPO}/issues/{issue_number}/comments"
+
+  try:
+    response = get_request(url)
+  except requests.exceptions.RequestException as e:
+    return error_response(f"Error: {e}")
+  return {"status": "success", "comments": response}
+
+
+root_agent = Agent(
+    model="gemini-2.5-pro",
+    name="adk_issue_formatting_assistant",
+    description="Check ADK issue format and content.",
+    instruction=f"""
+      # 1. IDENTITY
+      You are an AI assistant designed to help maintain the quality and consistency of issues in our GitHub repository.
+      Your primary role is to act as a "GitHub Issue Format Validator." You will analyze new and existing **open** issues
+      to ensure they contain all the necessary information as required by our templates. You are helpful, polite,
+      and precise in your feedback.
+
+      # 2. CONTEXT & RESOURCES
+      * **Repository:** You are operating on the GitHub repository `{OWNER}/{REPO}`.
+      * **Bug Report Template:** (`{BUG_REPORT_TEMPLATE}`)
+      * **Feature Request Template:** (`{FREATURE_REQUEST_TEMPLATE}`)
+
+      # 3. CORE MISSION
+      Your goal is to check if a GitHub issue, identified as either a "bug" or a "feature request,"
+      contains all the information required by the corresponding template. If it does not, your job is
+      to post a single, helpful comment asking the original author to provide the missing information.
+      {APPROVAL_INSTRUCTION}
+
+      **IMPORTANT NOTE:**
+      * You add one comment at most each time you are invoked.
+      * Don't proceed to other issues which are not the target issues.
+      * Don't take any action on closed issues.
+
+      # 4. BEHAVIORAL RULES & LOGIC
+
+      ## Step 1: Identify Issue Type & Applicability
+
+      Your first task is to determine if the issue is a valid target for validation.
+
+      1.  **Assess Content Intent:** You must perform a quick semantic check of the issue's title, body, and comments.
+          If you determine the issue's content is fundamentally *not* a bug report or a feature request
+          (for example, it is a general question, a request for help, or a discussion prompt), then you must ignore it.
+      2. **Exit Condition:** If the issue does not clearly fall into the categories of "bug" or "feature request"
+          based on both its labels and its content, **take no action**.
+
+      ## Step 2: Analyze the Issue Content
+
+      If you have determined the issue is a valid bug or feature request, your analysis depends on whether it has comments.
+
+      **Scenario A: Issue has NO comments**
+      1.  Read the main body of the issue.
+      2.  Compare the content of the issue body against the required headings/sections in the relevant template (Bug or Feature).
+      3.  Check for the presence of content under each heading. A heading with no content below it is considered incomplete.
+      4.  If one or more sections are missing or empty, proceed to Step 3.
+      5.  If all sections are filled out, your task is complete. Do nothing.
+
+      **Scenario B: Issue HAS one or more comments**
+      1.  First, analyze the main issue body to see which sections of the template are filled out.
+      2.  Next, read through **all** the comments in chronological order.
+      3.  As you read the comments, check if the information provided in them satisfies any of the template sections that were missing from the original issue body.
+      4.  After analyzing the body and all comments, determine if any required sections from the template *still* remain unaddressed.
+      5.  If one or more sections are still missing information, proceed to Step 3.
+      6.  If the issue body and comments *collectively* provide all the required information, your task is complete. Do nothing.
+
+      ## Step 3: Formulate and Post a Comment (If Necessary)
+
+      If you determined in Step 2 that information is missing, you must post a **single comment** on the issue.
+
+      Please include a bolded note in your comment that this comment was added by an ADK agent.
+
+      **Comment Guidelines:**
+      * **Be Polite and Helpful:** Start with a friendly tone.
+      * **Be Specific:** Clearly list only the sections from the template that are still missing. Do not list sections that have already been filled out.
+      * **Address the Author:** Mention the issue author by their username (e.g., `@username`).
+      * **Provide Context:** Explain *why* the information is needed (e.g., "to help us reproduce the bug" or "to better understand your request").
+      * **Do not be repetitive:** If you have already commented on an issue asking for information, do not comment again unless new information has been added and it's still incomplete.
+
+      **Example Comment for a Bug Report:**
+      > **Response from ADK Agent**
+      >
+      > Hello @[issue-author-username], thank you for submitting this issue!
+      >
+      > To help us investigate and resolve this bug effectively, could you please provide the missing details for the following sections of our bug report template:
+      >
+      > * **To Reproduce:** (Please provide the specific steps required to reproduce the behavior)
+      > * **Desktop (please complete the following information):** (Please provide OS, Python version, and ADK version)
+      >
+      > This information will give us the context we need to move forward. Thanks!
+
+      **Example Comment for a Feature Request:**
+      > **Response from ADK Agent**
+      >
+      > Hi @[issue-author-username], thanks for this great suggestion!
+      >
+      > To help our team better understand and evaluate your feature request, could you please provide a bit more information on the following section:
+      >
+      > * **Is your feature request related to a problem? Please describe.**
+      >
+      > We look forward to hearing more about your idea!
+
+      # 5. FINAL INSTRUCTION
+
+      Execute this process for the given GitHub issue. Your final output should either be **[NO ACTION]**
+      if the issue is complete or invalid, or **[POST COMMENT]** followed by the exact text of the comment you will post.
+
+      Please include your justification for your decision in your output.
+    """,
+    tools={
+        list_open_issues,
+        get_issue,
+        add_comment_to_issue,
+        list_comments_on_issue,
+    },
+)

contributing/samples/adk_issue_formatting_agent/settings.py

@@ -0,0 +1,33 @@
+# 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.
+
+import os
+
+from dotenv import load_dotenv
+
+load_dotenv(override=True)
+
+GITHUB_BASE_URL = "https://api.github.com"
+
+GITHUB_TOKEN = os.getenv("GITHUB_TOKEN")
+if not GITHUB_TOKEN:
+  raise ValueError("GITHUB_TOKEN environment variable not set")
+
+OWNER = os.getenv("OWNER", "google")
+REPO = os.getenv("REPO", "adk-python")
+EVENT_NAME = os.getenv("EVENT_NAME")
+ISSUE_NUMBER = os.getenv("ISSUE_NUMBER")
+ISSUE_COUNT_TO_PROCESS = os.getenv("ISSUE_COUNT_TO_PROCESS")
+
+IS_INTERACTIVE = os.environ.get("INTERACTIVE", "1").lower() in ["true", "1"]

contributing/samples/adk_issue_formatting_agent/utils.py

@@ -0,0 +1,53 @@
+# 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 typing import Any
+
+from adk_issue_formatting_agent.settings import GITHUB_TOKEN
+import requests
+
+headers = {
+    "Authorization": f"token {GITHUB_TOKEN}",
+    "Accept": "application/vnd.github.v3+json",
+}
+
+
+def get_request(
+    url: str, params: dict[str, Any] | None = None
+) -> dict[str, Any]:
+  if params is None:
+    params = {}
+  response = requests.get(url, headers=headers, params=params, timeout=60)
+  response.raise_for_status()
+  return response.json()
+
+
+def post_request(url: str, payload: Any) -> dict[str, Any]:
+  response = requests.post(url, headers=headers, json=payload, timeout=60)
+  response.raise_for_status()
+  return response.json()
+
+
+def error_response(error_message: str) -> dict[str, Any]:
+  return {"status": "error", "message": error_message}
+
+
+def read_file(file_path: str) -> str:
+  """Read the content of the given file."""
+  try:
+    with open(file_path, "r") as f:
+      return f.read()
+  except FileNotFoundError:
+    print(f"Error: File not found: {file_path}.")
+    return ""

contributing/samples/artifact_save_text/agent.py

@@ -31,7 +31,7 @@ root_agent = Agent(
     model='gemini-2.0-flash',
     name='log_agent',
     description='Log user query.',
-    instruction="""Always log the user query and reploy "kk, I've logged."
+    instruction="""Always log the user query and reply "kk, I've logged."
     """,
     tools=[log_query],
     generate_content_config=types.GenerateContentConfig(

contributing/samples/bigquery/agent.py

@@ -60,7 +60,7 @@ bigquery_toolset = BigQueryToolset(
 # debug CLI
 root_agent = llm_agent.Agent(
     model="gemini-2.0-flash",
-    name="hello_agent",
+    name="bigquery_agent",
     description=(
         "Agent to answer questions about BigQuery data and models and execute"
         " SQL queries."

contributing/samples/google_api/README.md

@@ -1,45 +1,40 @@
-# BigQuery Sample
+# Google API Tools Sample
 
 ## Introduction
 
-This sample tests and demos the BigQuery support in ADK via two tools:
+This sample tests and demos Google API tools available in the
+`google.adk.tools.google_api_tool` module. We pick the following BigQuery API
+tools for this sample agent:
 
-* 1. bigquery_datasets_list:
+1. `bigquery_datasets_list`: List user's datasets.
 
-    List user's datasets.
+2. `bigquery_datasets_get`: Get a dataset's details.
 
-* 2. bigquery_datasets_get:
-    Get a dataset's details.
+3. `bigquery_datasets_insert`: Create a new dataset.
 
-* 3. bigquery_datasets_insert:
-    Create a new dataset.
+4. `bigquery_tables_list`: List all tables in a dataset.
 
-* 4. bigquery_tables_list:
-    List all tables in a dataset.
+5. `bigquery_tables_get`: Get a table's details.
 
-* 5. bigquery_tables_get:
-    Get a table's details.
-
-* 6. bigquery_tables_insert:
-    Insert a new table into a dataset.
+6. `bigquery_tables_insert`: Insert a new table into a dataset.
 
 ## How to use
 
-* 1. Follow https://developers.google.com/identity/protocols/oauth2#1.-obtain-oauth-2.0-credentials-from-the-dynamic_data.setvar.console_name. to get your client id and client secret.
+1. Follow https://developers.google.com/identity/protocols/oauth2#1.-obtain-oauth-2.0-credentials-from-the-dynamic_data.setvar.console_name. to get your client id and client secret.
   Be sure to choose "web" as your client type.
 
-* 2. Configure your `.env` file to add two variables:
+2. Configure your `.env` file to add two variables:
 
   * OAUTH_CLIENT_ID={your client id}
   * OAUTH_CLIENT_SECRET={your client secret}
 
   Note: don't create a separate `.env` file , instead put it to the same `.env` file that stores your Vertex AI or Dev ML credentials
 
-* 3. Follow https://developers.google.com/identity/protocols/oauth2/web-server#creatingcred to add http://localhost/dev-ui/ to "Authorized redirect URIs".
+3. Follow https://developers.google.com/identity/protocols/oauth2/web-server#creatingcred to add http://localhost/dev-ui/ to "Authorized redirect URIs".
 
   Note: localhost here is just a hostname that you use to access the dev ui, replace it with the actual hostname you use to access the dev ui.
 
-* 4. For 1st run, allow popup for localhost in Chrome.
+4. For 1st run, allow popup for localhost in Chrome.
 
 ## Sample prompt
 

contributing/samples/google_api/__init__.py

@@ -0,0 +1,15 @@
+# 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 . import agent

contributing/samples/google_api/agent.py

@@ -40,7 +40,7 @@ bigquery_toolset = BigQueryToolset(
 
 root_agent = Agent(
     model="gemini-2.0-flash",
-    name="bigquery_agent",
+    name="google_api_bigquery_agent",
     instruction="""
       You are a helpful Google BigQuery agent that help to manage users' data on Google BigQuery.
       Use the provided tools to conduct various operations on users' data in Google BigQuery.

contributing/samples/hello_world_litellm_add_function_to_prompt/__init__.py

@@ -0,0 +1,16 @@
+# 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 . import agent

contributing/samples/hello_world_litellm_add_function_to_prompt/agent.py

@@ -0,0 +1,78 @@
+# 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.
+
+
+import random
+
+from google.adk import Agent
+from google.adk.models.lite_llm import LiteLlm
+from langchain_core.utils.function_calling import convert_to_openai_function
+
+
+def roll_die(sides: int) -> int:
+  """Roll a die and return the rolled result.
+
+  Args:
+    sides: The integer number of sides the die has.
+
+  Returns:
+    An integer of the result of rolling the die.
+  """
+  return random.randint(1, sides)
+
+
+def check_prime(number: int) -> str:
+  """Check if a given number is prime.
+
+  Args:
+    number: The input number to check.
+
+  Returns:
+    A str indicating the number is prime or not.
+  """
+  if number <= 1:
+    return f"{number} is not prime."
+  is_prime = True
+  for i in range(2, int(number**0.5) + 1):
+    if number % i == 0:
+      is_prime = False
+      break
+  if is_prime:
+    return f"{number} is prime."
+  else:
+    return f"{number} is not prime."
+
+
+root_agent = Agent(
+    model=LiteLlm(
+        model="vertex_ai/meta/llama-4-maverick-17b-128e-instruct-maas",
+        # If the model is not trained with functions and you would like to
+        # enable function calling, you can add functions to the models, and the
+        # functions will be added to the prompts during inferences.
+        functions=[
+            convert_to_openai_function(roll_die),
+            convert_to_openai_function(check_prime),
+        ],
+    ),
+    name="data_processing_agent",
+    description="""You are a helpful assistant.""",
+    instruction="""
+      You are a helpful assistant, and call tools optionally.
+      If call tools, the tool format should be in json, and the tool arguments should be parsed from users inputs.
+    """,
+    tools=[
+        roll_die,
+        check_prime,
+    ],
+)

contributing/samples/hello_world_litellm_add_function_to_prompt/main.py

@@ -0,0 +1,81 @@
+# 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.
+
+
+import asyncio
+import time
+
+import agent
+from dotenv import load_dotenv
+from google.adk import Runner
+from google.adk.artifacts import InMemoryArtifactService
+from google.adk.cli.utils import logs
+from google.adk.sessions import InMemorySessionService
+from google.adk.sessions import Session
+from google.genai import types
+
+load_dotenv(override=True)
+logs.log_to_tmp_folder()
+
+
+async def main():
+  app_name = 'my_app'
+  user_id_1 = 'user1'
+  session_service = InMemorySessionService()
+  artifact_service = InMemoryArtifactService()
+  runner = Runner(
+      app_name=app_name,
+      agent=agent.root_agent,
+      artifact_service=artifact_service,
+      session_service=session_service,
+  )
+  session_11 = await session_service.create_session(
+      app_name=app_name, user_id=user_id_1
+  )
+
+  async def run_prompt(session: Session, new_message: str):
+    content = types.Content(
+        role='user', parts=[types.Part.from_text(text=new_message)]
+    )
+    print('** User says:', content.model_dump(exclude_none=True))
+    async for event in runner.run_async(
+        user_id=user_id_1,
+        session_id=session.id,
+        new_message=content,
+    ):
+      if event.content.parts:
+        part = event.content.parts[0]
+        if part.text:
+          print(f'** {event.author}: {part.text}')
+        if part.function_call:
+          print(f'** {event.author} calls tool: {part.function_call}')
+        if part.function_response:
+          print(
+              f'** {event.author} gets tool response: {part.function_response}'
+          )
+
+  start_time = time.time()
+  print('Start time:', start_time)
+  print('------------------------------------')
+  await run_prompt(session_11, 'Hi, introduce yourself.')
+  await run_prompt(session_11, 'Roll a die with 100 sides.')
+  await run_prompt(session_11, 'Check if it is prime.')
+  end_time = time.time()
+  print('------------------------------------')
+  print('End time:', end_time)
+  print('Total time:', end_time - start_time)
+
+
+if __name__ == '__main__':
+  asyncio.run(main())

contributing/samples/oauth_calendar_agent/agent.py

@@ -13,7 +13,6 @@
 # limitations under the License.
 
 from datetime import datetime
-import json
 import os
 
 from dotenv import load_dotenv
@@ -27,8 +26,8 @@ from google.adk.auth import AuthCredential
 from google.adk.auth import AuthCredentialTypes
 from google.adk.auth import OAuth2Auth
 from google.adk.tools import ToolContext
+from google.adk.tools.authenticated_function_tool import AuthenticatedFunctionTool
 from google.adk.tools.google_api_tool import CalendarToolset
-from google.auth.transport.requests import Request
 from google.oauth2.credentials import Credentials
 from googleapiclient.discovery import build
 
@@ -56,6 +55,7 @@ def list_calendar_events(
     end_time: str,
     limit: int,
     tool_context: ToolContext,
+    credential: AuthCredential,
 ) -> list[dict]:
   """Search for calendar events.
 
@@ -80,84 +80,11 @@ def list_calendar_events(
   Returns:
       list[dict]: A list of events that match the search criteria.
   """
-  creds = None
 
-  # Check if the tokes were already in the session state, which means the user
-  # has already gone through the OAuth flow and successfully authenticated and
-  # authorized the tool to access their calendar.
-  if "calendar_tool_tokens" in tool_context.state:
-    creds = Credentials.from_authorized_user_info(
-        tool_context.state["calendar_tool_tokens"], SCOPES
-    )
-  if not creds or not creds.valid:
-    # If the access token is expired, refresh it with the refresh token.
-    if creds and creds.expired and creds.refresh_token:
-      creds.refresh(Request())
-    else:
-      auth_scheme = OAuth2(
-          flows=OAuthFlows(
-              authorizationCode=OAuthFlowAuthorizationCode(
-                  authorizationUrl="https://accounts.google.com/o/oauth2/auth",
-                  tokenUrl="https://oauth2.googleapis.com/token",
-                  scopes={
-                      "https://www.googleapis.com/auth/calendar": (
-                          "See, edit, share, and permanently delete all the"
-                          " calendars you can access using Google Calendar"
-                      )
-                  },
-              )
-          )
-      )
-      auth_credential = AuthCredential(
-          auth_type=AuthCredentialTypes.OAUTH2,
-          oauth2=OAuth2Auth(
-              client_id=oauth_client_id, client_secret=oauth_client_secret
-          ),
-      )
-      # If the user has not gone through the OAuth flow before, or the refresh
-      # token also expired, we need to ask users to go through the OAuth flow.
-      # First we check whether the user has just gone through the OAuth flow and
-      # Oauth response is just passed back.
-      auth_response = tool_context.get_auth_response(
-          AuthConfig(
-              auth_scheme=auth_scheme, raw_auth_credential=auth_credential
-          )
-      )
-      if auth_response:
-        # ADK exchanged the access token already for us
-        access_token = auth_response.oauth2.access_token
-        refresh_token = auth_response.oauth2.refresh_token
-
-        creds = Credentials(
-            token=access_token,
-            refresh_token=refresh_token,
-            token_uri=auth_scheme.flows.authorizationCode.tokenUrl,
-            client_id=oauth_client_id,
-            client_secret=oauth_client_secret,
-            scopes=list(auth_scheme.flows.authorizationCode.scopes.keys()),
-        )
-      else:
-        # If there are no auth response which means the user has not gone
-        # through the OAuth flow yet, we need to ask users to go through the
-        # OAuth flow.
-        tool_context.request_credential(
-            AuthConfig(
-                auth_scheme=auth_scheme,
-                raw_auth_credential=auth_credential,
-            )
-        )
-        # The return value is optional and could be any dict object. It will be
-        # wrapped in a dict with key as 'result' and value as the return value
-        # if the object returned is not a dict. This response will be passed
-        # to LLM to generate a user friendly message. e.g. LLM will tell user:
-        # "I need your authorization to access your calendar. Please authorize
-        # me so I can check your meetings for today."
-        return "Need User Authorization to access their calendar."
-    # We store the access token and refresh token in the session state for the
-    # next runs. This is just an example. On production, a tool should store
-    # those credentials in some secure store or properly encrypt it before store
-    # it in the session state.
-    tool_context.state["calendar_tool_tokens"] = json.loads(creds.to_json())
+  creds = Credentials(
+      token=credential.oauth2.access_token,
+      refresh_token=credential.oauth2.refresh_token,
+  )
 
   service = build("calendar", "v3", credentials=creds)
   events_result = (
@@ -208,6 +135,33 @@ root_agent = Agent(
 
       Currnet time: {_time}
 """,
-    tools=[list_calendar_events, calendar_toolset],
+    tools=[
+        AuthenticatedFunctionTool(
+            func=list_calendar_events,
+            auth_config=AuthConfig(
+                auth_scheme=OAuth2(
+                    flows=OAuthFlows(
+                        authorizationCode=OAuthFlowAuthorizationCode(
+                            authorizationUrl=(
+                                "https://accounts.google.com/o/oauth2/auth"
+                            ),
+                            tokenUrl="https://oauth2.googleapis.com/token",
+                            scopes={
+                                "https://www.googleapis.com/auth/calendar": "",
+                            },
+                        )
+                    )
+                ),
+                raw_auth_credential=AuthCredential(
+                    auth_type=AuthCredentialTypes.OAUTH2,
+                    oauth2=OAuth2Auth(
+                        client_id=oauth_client_id,
+                        client_secret=oauth_client_secret,
+                    ),
+                ),
+            ),
+        ),
+        calendar_toolset,
+    ],
     before_agent_callback=update_time,
 )

pyproject.toml

@@ -25,6 +25,7 @@ classifiers = [ # List of https://pypi.org/classifiers/
 ]
 dependencies = [
   # go/keep-sorted start
+  "PyYAML>=6.0.2",                                  # For APIHubToolset.
   "anyio>=4.9.0;python_version>='3.10'",            # For MCP Session Manager
   "authlib>=1.5.1",                                 # For RestAPI Tool
   "click>=8.1.8",                                   # For CLI tools
@@ -34,7 +35,7 @@ dependencies = [
   "google-cloud-secret-manager>=2.22.0",            # Fetching secrets in RestAPI Tool
   "google-cloud-speech>=2.30.0",                    # For Audio Transcription
   "google-cloud-storage>=2.18.0, <3.0.0",           # For GCS Artifact service
-  "google-genai>=1.17.0",                           # Google GenAI SDK
+  "google-genai>=1.21.1",                           # Google GenAI SDK
   "graphviz>=0.20.2",                               # Graphviz for graph rendering
   "mcp>=1.8.0;python_version>='3.10'",              # For MCP Toolset
   "opentelemetry-api>=1.31.0",                      # OpenTelemetry
@@ -43,7 +44,6 @@ dependencies = [
   "pydantic>=2.0, <3.0.0",                          # For data validation/models
   "python-dateutil>=2.9.0.post0",                   # For Vertext AI Session Service
   "python-dotenv>=1.0.0",                           # To manage environment variables
-  "PyYAML>=6.0.2",                                  # For APIHubToolset.
   "requests>=2.32.4",
   "sqlalchemy>=2.0",                                # SQL database ORM
   "starlette>=0.46.2",                              # For FastAPI CLI
@@ -70,9 +70,9 @@ dev = [
   # go/keep-sorted start
   "flit>=3.10.0",
   "isort>=6.0.0",
+  "mypy>=1.15.0",
   "pyink>=24.10.0",
   "pylint>=2.6.0",
-  "mypy>=1.15.0",
   # go/keep-sorted end
 ]
 
@@ -87,6 +87,7 @@ eval = [
   "google-cloud-aiplatform[evaluation]>=1.87.0",
   "pandas>=2.2.3",
   "tabulate>=0.9.0",
+  "rouge-score>=0.1.2",
   # go/keep-sorted end
 ]
 
@@ -97,7 +98,6 @@ test = [
   "langgraph>=0.2.60",               # For LangGraphAgent
   "litellm>=1.71.2",                 # For LiteLLM tests
   "llama-index-readers-file>=0.4.0", # For retrieval tests
-
   "pytest-asyncio>=0.25.0",
   "pytest-mock>=3.14.0",
   "pytest-xdist>=3.6.1",