ahttp-client
ahttp-client is a decorator-based HTTP client framework
that maps typed function parameters to HTTP requests.
Key Features
- Declare HTTP endpoints using
@requestor the@get,@post,@put,@patch,@delete, and@optionsdecoration methods. - Use
typing.Annotatedto set HTTP parameters such as the path, query, header, or body values. - Serialize typed request models and deserialize responses using registered codecs.
- Retry failed requests with configurable exception filters and exponential backoff.
- Customize the request lifecycle using
before_hookandafter_hookdecorators. - Swap between aiohttp, httpx, and requests without changing how a service is declared.
Installation
Install the extra for the HTTP client library you want to use. Python 3.11 or later is required.
pip install "ahttp-client[aiohttp]"
pip install "ahttp-client[httpx]"
pip install "ahttp-client[requests]"
Include the pydantic extra to serialize and deserialize Pydantic models.
pip install "ahttp-client[aiohttp,pydantic]"
Quick start
| Style | Supported client classes | Session class |
|---|---|---|
| Async | aiohttp.ClientSession, httpx.AsyncClient |
AsyncSession |
| Sync | requests.Session, httpx.Client |
Session |
Asynchronous Client
Declare a service by extending AsyncSession, then decorate coroutine methods
with an HTTP method and path. Annotated parameters determine where values are
placed in the request.
import asyncio
from typing import Annotated, Any
import aiohttp
from ahttp_client import AsyncSession, Path, Response, get
class GitHubService(AsyncSession):
def __init__(self):
super().__init__("https://api.github.com", aiohttp.ClientSession)
@get("/users/{user}/repos")
async def list_repositories(
self, response: Response, user: Annotated[str, Path]
) -> list[dict[str, Any]]:
return response.json()
async def main():
async with GitHubService() as service:
repositories = await service.list_repositories(user="gunyu1019")
print(repositories)
asyncio.run(main())
AsyncSession closes its underlying HTTP client when the async with block
ends. Decorated responses are also closed automatically after the handler
returns.
Synchronous Client
Use Session and a regular function with requests.Session or httpx.Client.
from typing import Annotated, Any
import requests
from ahttp_client import Path, Response, Session, get
class GitHubService(Session):
def __init__(self):
super().__init__("https://api.github.com", requests.Session)
@get("/users/{user}/repos")
def list_repositories(
self, response: Response, user: Annotated[str, Path]
) -> list[dict[str, Any]]:
return response.json()
with GitHubService() as service:
repositories = service.list_repositories(user="gunyu1019")
print(repositories)
Request components
Use Annotated to describe dynamic request values.
from typing import Annotated
from ahttp_client import BodyJson, Header, Path, Query, Response, post
class UserService(AsyncSession):
@post("/users/{user_id}")
async def update_user(
self,
response: Response,
user_id: Annotated[int, Path],
verbose: Annotated[bool, Query],
authorization: Annotated[str, Header.custom_name("Authorization")],
display_name: Annotated[str, BodyJson.custom_key("profile.displayName")],
) -> dict:
return response.json()
| Component | Request location |
|---|---|
Path |
A {placeholder} in the path |
Query |
Query string |
Header |
Request header |
BodyJson |
JSON body field; supports nested keys |
BodyForm |
URL-encoded or multipart form field |
Body |
Complete raw or JSON request body |
Set directly_response=True on a request (or a session) when you need the
Response object itself instead of running the decorated handler. In that
case, close it yourself with await response.async_close() for async clients
or response.close() for sync clients.
Set raise_on=True when an endpoint must treat every status other than 200
as a failure. It raises HTTPException before the response is returned or
deserialized, so it also participates in configured retry handling.
@get("/health", raise_on=True)
async def health(self) -> None:
...
Model serialization
Registered codecs can convert a complete Body parameter before transport and
validate a direct response from its return annotation. When Pydantic is
installed, BaseModel types and nested model containers are supported
automatically.
Use @serialize and @deserialize to pass codec options. If the model argument
is omitted, the request body and return annotations select the codec after the
request decorator is applied.
from typing import Annotated
from pydantic import BaseModel
from ahttp_client import AsyncSession, Body, post
from ahttp_client.serializer import deserialize, serialize
class CreateUser(BaseModel):
name: str
nickname: str | None = None
class User(BaseModel):
id: int
name: str
class UserService(AsyncSession):
@post("/users", directly_response=True)
@serialize(exclude_none=True)
@deserialize(strict=True)
async def create_user(
self,
user: Annotated[CreateUser, Body],
) -> User:
...
In this example, the request body is produced with
BaseModel.model_dump(mode="json", exclude_none=True), and the JSON response is
validated as User. Because directly_response=True selects deserialized mode
from the registered return type, the decorated method body is not executed.
Pass a model explicitly, such as @serialize(CreateUser), when it cannot be
inferred from an annotation.
Static type checking
The package includes an optional mypy plugin for declarative endpoint methods.
It preserves their public call signatures and allows a skipped direct-response
body to contain only ... or pass, including when mypy strict mode is used.
No additional package is required:
[mypy]
plugins = ahttp_client.mypy
When Pydantic models are also checked, both installed plugins can be enabled:
[mypy]
plugins = pydantic.mypy, ahttp_client.mypy
Retries
Use @retry to repeat a request when a selected exception is raised. By
default, HTTPServerError retries HTTP 5xx failures up to three times after
the initial request. Call Response.raise_for_status() from after_request()
when HTTP error responses should participate in retry handling.
Alternatively, use retry_on_status to retry returned response statuses
directly. This does not require an exception or an after_request() hook;
responses from attempts that will be retried are closed automatically.
from typing import Annotated, Any
from ahttp_client import (
AsyncSession,
HTTPServerError,
Path,
Response,
get,
retry,
)
class GitHubService(AsyncSession):
async def after_request(self, response: Response) -> Response:
response.raise_for_status()
return response
@retry(
max_retries=3,
backoff_factor=0.5,
retry_on=(HTTPServerError, TimeoutError),
retry_on_status=(502, 503, 504),
max_delay=4.0,
)
@get("/users/{user}/repos")
async def list_repositories(
self, response: Response, user: Annotated[str, Path]
) -> list[dict[str, Any]]:
return response.json()
The wait before retry attempt n is
backoff_factor * 2 ** (n - 1) seconds and is capped by max_delay when set.
Pass one exception class or a tuple through retry_on to include transport or
application failures. Pass an HTTP status code or a tuple through
retry_on_status to retry responses without raising an exception. If its
retry budget is exhausted, the last response is returned normally.
Exceptions raised during request transport or the session-level
after_request() hook are eligible for retry. Request-level after_hook
callbacks run after the retry operation and are not retried. Retry counts and
delays must be finite, non-negative values.
Retries are enabled automatically only for idempotent HTTP methods. Retrying a
POST, PATCH, or another non-idempotent request can duplicate a server-side
operation, so it requires an explicit retry_unsafe=True opt-in:
@retry(max_retries=1, retry_unsafe=True)
@post("/jobs")
async def create_job(self, payload: Annotated[dict[str, Any], BodyJson]) -> None:
...
Hooks
Attach a hook to a decorated request to modify it before dispatch or transform its result afterward. Async requests require async hooks; sync requests require regular functions.
class GitHubService(AsyncSession):
@get("/user")
async def current_user(self, response: Response) -> dict:
return response.json()
@current_user.before_hook
async def add_authorization(self, request, path):
request.headers["Authorization"] = "Bearer <token>"
return request, path
Override before_request() or after_request() on AsyncSession or Session
to apply the same behavior to every request in a service.
Documentation
Metadata
Release files for ahttp-client 2.0.0
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | Actions |
|---|---|---|---|---|
| ahttp_client-2.0.0.tar.gz | 49.6 kB | With provenance | Details for ahttp_client-2.0.0.tar.gz |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Actions Reset |
|---|---|---|---|---|
| ahttp_client-2.0.0-py3-none-any.whl | Python 3 | none | any | Details for ahttp_client-2.0.0-py3-none-any.whl |
Total release size: 120.2 kB
Release files / ahttp_client-2.0.0.tar.gz
| Download URL | ahttp_client-2.0.0.tar.gz |
|---|---|
| Size | 49.6 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
6f65d98ac890fa5d223f68083f46e03111888d64ee48349d2766d89c6879a618
|
|
BLAKE2b-256 checksum How to use checksums |
33931fd122aa91a9d6017064d7ca000f675943dd3a90d3a6ee4984e6a7ed098c
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/6.1.0 CPython/3.12.8
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Jul 25, 2026.
Transparency logRelease files / ahttp_client-2.0.0-py3-none-any.whl
| Download URL | ahttp_client-2.0.0-py3-none-any.whl |
|---|---|
| Size | 70.5 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
bac2df86f5edce9b2f28ab1bd96481f647bb67433db16598c428107818b0bfac
|
|
BLAKE2b-256 checksum How to use checksums |
c6440c303a4b338c4ea9cea68859b52ca6448ce3480aefcbaa456ca3f8424a60
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/6.1.0 CPython/3.12.8
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Jul 25, 2026.
Transparency logSecurity for ahttp-client 2.0.0
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
Full provenance
Both files in this release have been uploaded with provenance metadata.
Attestation for ahttp_client-2.0.0.tar.gz
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Jul 25, 2026.
Transparency logA note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.