-
Notifications
You must be signed in to change notification settings - Fork 161
feat: Adding stand-alone support for RESTful API serving #297
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
base: restful
Are you sure you want to change the base?
Conversation
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Summary of Changes
Hello @pstephengoogle, I'm Gemini Code Assist1! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed!
This pull request introduces the capability to serve the A2A RESTful API as a standalone service, offering an alternative to gRPC-based serving with Envoy transcoding. It achieves this by adding new application and request handler components that directly process REST (JSON+HTTP) requests, map them to the core A2A logic, and manage responses, including Server-Sent Events.
Highlights
- Standalone REST API Support: Introduces new
RESTApplication
andA2ARESTFastAPIApplication
classes to enable serving the A2A RESTful API independently, without reliance on gRPC or Envoy transcoding. - REST Request Handling: Adds a dedicated
RESTHandler
that maps incoming REST (JSON+HTTP) requests to the appropriate underlying A2A request handler methods, ensuring consistency with existing gRPC service definitions by leveraging protobuf for data serialization/deserialization. - FastAPI Integration: Implements
A2ARESTFastAPIApplication
to build a FastAPI application, providing concrete endpoints for A2A operations such as message sending (including streaming), task retrieval, and push notification management.
Using Gemini Code Assist
The full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips.
Invoking Gemini
You can request assistance from Gemini at any point in your pull request via creating an issue comment (i.e. comment on the pull request page) using either /gemini <command>
or @gemini-code-assist <command>
. Below is a summary of the supported commands.
Feature | Command | Description |
---|---|---|
Code Review | /gemini review |
Performs a code review for the current pull request in its current state. |
Pull Request Summary | /gemini summary |
Provides a summary of the current pull request in its current state. |
Comment | @gemini-code-assist | Responds in comments when explicitly tagged, both in issue comments and review comments. |
Help | /gemini help |
Displays a list of available commands. |
Customization
To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a .gemini/
folder in the base of the repository. Detailed instructions can be found here.
Limitations & Feedback
Gemini Code Assist is currently in preview and may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with 👍 and 👎 on @gemini-code-assist comments to provide feedback.
You can also get AI-powered code generation, chat, as well as code reviews directly in the IDE at no cost with the Gemini Code Assist IDE Extension.
Footnotes
-
Review the Privacy Notices, Generative AI Prohibited Use Policy, Terms of Service, and learn how to configure Gemini Code Assist in GitHub here. Gemini can make mistakes, so double check it and use code with caution. ↩
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Code Review
This pull request introduces stand-alone support for serving the RESTful API. Several critical and high-severity issues concerning attribute access, routing logic, error handling, and type safety were identified.
# Description Push notifications are not being triggered for message/send operations without an initial task.
# Description This update improves the robustness of the `new_task` function by adding explicit error handling for invalid `Message` objects. Previously, `new_task` did not validate its input, which could lead to unexpected `TypeError` or `ValueError` exceptions when a malformed `Message` was passed. This change introduces checks to ensure that a `Message` has a valid `role` and a non-empty `parts` list before a `Task` is created. Additionally, this update includes corresponding unit tests to verify the new error handling and ensures that the correct exceptions are raised for invalid inputs.
# Description Refactored `tests/client/test_auth_middleware.py` and added test cases to improve coverage for `credentials.py` and `interceptor.py`. - `src/a2a/client/auth/credentials.py`: 94% -> 100% - `src/a2a/client/auth/interceptor.py`: 71% -> 94% Signed-off-by: Shingo OKAWA <shingo.okawa.g.h.c@gmail.com>
🤖 I have created a release *beep* *boop* --- ## [0.2.12](v0.2.11...v0.2.12) (2025-07-14) ### Features * add `metadata` property to `RequestContext` ([#302](#302)) ([e781ced](e781ced)) * add A2ABaseModel ([#292](#292)) ([24f2eb0](24f2eb0)) * add support for notification tokens in PushNotificationSender ([#266](#266)) ([75aa4ed](75aa4ed)) * Update A2A types from specification 🤖 ([#289](#289)) ([ecb321a](ecb321a)) ### Bug Fixes * add proper a2a request body documentation to Swagger UI ([#276](#276)) ([4343be9](4343be9)), closes [#274](#274) * Handle asyncio.cancellederror and raise to propagate back ([#293](#293)) ([9d6cb68](9d6cb68)) * Improve error handling in task creation ([#294](#294)) ([6412c75](6412c75)) * Resolve dependency issue with sql stores ([#303](#303)) ([2126828](2126828)) * Send push notifications for message/send ([#298](#298)) ([0274112](0274112)) * **server:** Improve event consumer error handling ([#282](#282)) ([a5786a1](a5786a1)) --- This PR was generated with [Release Please](https://github.com/googleapis/release-please). See [documentation](https://github.com/googleapis/release-please#release-please). Co-authored-by: release-please[bot] <55107282+release-please[bot]@users.noreply.github.com>
# Description Introduces support for Database backed Push Notification Config Store. Push Notification Config persisted to the database table gets encrypted when encryption_key param is set for the DatabasePushNotificationConfigStore. Usage: ```py httpx_client = httpx.AsyncClient() engine = create_async_engine( "sqlite+aiosqlite:///file::memory:?cache=shared", echo=False ) push_config_store = DatabasePushNotificationConfigStore(engine=engine, encryption_key=Fernet.generate_key()) push_sender = BasePushNotificationSender( httpx_client=httpx_client, config_store=push_config_store) request_handler = DefaultRequestHandler( agent_executor=CurrencyAgentExecutor(), task_store=DatabaseTaskStore(engine=engine), push_config_store=push_config_store, push_sender=push_sender) ```
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Can you also add tests for these new handlers?
|
||
def build( | ||
self, | ||
agent_card_url: str = '/.well-known/agent.json', |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
agent_card_url: str = '/.well-known/agent.json', | |
agent_card_url: str = AGENT_CARD_WELL_KNOWN_PATH, |
Also add import:
from a2a.utils.constants import AGENT_CARD_WELL_KNOWN_PATH
logger = logging.getLogger(__name__) | ||
|
||
|
||
class RESTApplication: |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Should this be called A2ARESTApplication
?
This update addresses an issue where the `completed_task` function in `a2a/utils/task.py` did not perform any validation on the `artifacts` list. This could lead to unexpected behavior if an empty or invalid list was provided. This change introduces a validation check to ensure that the `artifacts` list is a non-empty list of `Artifact` objects, raising a `ValueError` if the validation fails. **Changes:** - Modified `a2a/utils/task.py` to add a validation check for the `artifacts` parameter in the `completed_task` function. - Updated `tests/utils/test_task.py` to include tests for the new validation logic, covering cases with empty lists and lists containing invalid items.
def __init__( | ||
self, | ||
agent_card: AgentCard, | ||
http_handler: RequestHandler, |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
nit: this param should be called request_handler, right? As it is the base request handler, there's nothing related to HTTP.
) | ||
return JSONResponse( | ||
'{"message": ' + error.message + '}', | ||
status_code=404, |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This would not be 404. Default value would be 500. But we should check for error and attach 4XX or 5XX HTTP error codes.
Not sure why, but this PR seems to have Krishna's persistence store changes as well. |
|
||
def routes(self) -> dict[tuple[str, str], Callable[[Request], Any]]: | ||
routes = { | ||
('/v1/message:send', 'POST'): ( |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Might want to check with Google AIP readability.
Since this is effectively creating a new message, do we need the ":send" custom operation. Can't we just call it ('/v1/message', 'POST')
Adding stand-alone support for RESTful API serving
Though the RESTful API (JSON+HTTP) can be served from an envoy transcoding of the gRPC service that may not always make sense in all cases. Instead, this PR also adds the ability to serve the RESTful API stand-alone, without needing to also setup the gRPC service.