Skip to main content

Thin wrapper around FastAPI

Project description

asapi

A thin opinionated wrapper around FastAPI. Because it's wrapping FastAPI you can work it into your existing projects.

Explicit composition root

FastAPI uses callbacks inside of Depends to do it's dependency injection. This forces you to end up using multiple layers of Depends to compose your application. The creation of these Depends resources often ends up distributed across modules so it's hard to know where something is initialized.

FastAPI also has no application-level dependencies, so you end up having to use globals to share resources across requests.

asapi solves this by having an explicit composition root where you can define all your dependencies in one place.

Endpoints then use Injected[DependencyType] to get access to the dependencies they need.

Example

from __future__ import annotations

import anyio
from fastapi import FastAPI
from psycopg_pool import AsyncConnectionPool
from asapi import FromPath, Injected, serve, bind


app = FastAPI()


@app.get("/hello/{name}")
async def hello(
    name: FromPath[str],
    pool: Injected[AsyncConnectionPool],
) -> str:
    async with pool.connection() as conn:
        async with conn.cursor() as cur:
            await cur.execute("SELECT '¡Hola ' || %(name)s || '!'", {"name": name})
            res = await cur.fetchone()
            assert res is not None
            return res[0]

TODO: in the future I'd like to provide a wrapper around APIRouter and FastAPI that also forces you to mark every argument to an endpoint as Injected, Query, Path, Body, which makes it explicit where arguments are coming from with minimal boilerplate.

Run in your event loop

FastAPI recommends using Uvicorn to run your application (note: if you're using Gunicorn you probably don't need to unless you're deploying on a a 'bare meta' server with multiple cores like a large EC2 instance).

But using uvicorn app:app from the command line has several issues:

  1. It takes control of the event loop and startup out of your hands. You have to rely on Uvicorn to configure the event loop, configure logging, etc.
  2. You'll have to use ASGI lifespans to initialize your resources, or the globals trick mentioned above.
  3. You can't run anything else in the event loop (e.g. a background worker).

asapi solves this by providing a serve function that you can use to run your application in your own event loop.

from __future__ import annotations

import anyio
from asapi import serve
from fastapi import FastAPI

app = FastAPI()

@app.get("/")
async def root() -> dict[str, str]:
    return {"message": "Hello World"}


async def main():
    await serve(app, 8000)

if __name__ == "__main__":
    anyio.run(main)

Now you have full control of the event loop and can make database connections, run background tasks, etc. Combined with the explicit composition root, you can initialize all your resources in one place, bind them to an application instance that is specific to this event loop and inject them into the endpoints that need them, all without global state or multiple layers of Depends.

from __future__ import annotations

import logging
from typing import Any
import anyio
from fastapi import FastAPI, APIRouter
from psycopg_pool import AsyncConnectionPool
from asapi import FromPath, Injected, serve, bind


router = APIRouter()


@router.get("/hello/{name}")
async def hello(name: FromPath[str], pool: Injected[AsyncConnectionPool]) -> str:
    async with pool.connection() as conn:
        async with conn.cursor() as cur:
            await cur.execute("SELECT '¡Hola ' || %(name)s || '!'", {"name": name})
            res = await cur.fetchone()
            assert res is not None
            return res[0]


def create_app(pool: AsyncConnectionPool[Any]) -> FastAPI:
    app = FastAPI()
    bind(app, AsyncConnectionPool, pool)
    app.include_router(router)
    return app


async def main() -> None:
    logging.basicConfig(level=logging.INFO)

    async with AsyncConnectionPool(
        "postgres://postgres:postgres@localhost:54320/postgres"
    ) as pool:
        app = create_app(pool)
        await serve(app, 9000)


if __name__ == "__main__":
    anyio.run(main)

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

asapi-0.7.0.tar.gz (49.1 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

asapi-0.7.0-py3-none-any.whl (6.4 kB view details)

Uploaded Python 3

File details

Details for the file asapi-0.7.0.tar.gz.

File metadata

  • Download URL: asapi-0.7.0.tar.gz
  • Upload date:
  • Size: 49.1 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: uv/0.5.1

File hashes

Hashes for asapi-0.7.0.tar.gz
Algorithm Hash digest
SHA256 4dcfa9eebef5bdb4c40786f3aab681a9d7de900a4e6ca1eedaa0dfec19eba251
MD5 869b2465e6dbab1f0398a4717696415b
BLAKE2b-256 3aa326596c00e320f4e2e72263fd323fc123a4fc7f30d92e3230db69778bbd73

See more details on using hashes here.

File details

Details for the file asapi-0.7.0-py3-none-any.whl.

File metadata

  • Download URL: asapi-0.7.0-py3-none-any.whl
  • Upload date:
  • Size: 6.4 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: uv/0.5.1

File hashes

Hashes for asapi-0.7.0-py3-none-any.whl
Algorithm Hash digest
SHA256 36b32d6ccfc494633cdb7c3b55c4e49b2514951a44ce7da759f4a92bf7941a19
MD5 4bfb6b70c86a4a168b2fb80e6e21f1fd
BLAKE2b-256 1b2da72a93deb46c23546e4b20fb2f5473399458651d892542773f67f14ebf2c

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page