Aiohttp pydantic - Aiohttp View to validate and parse request

Aiohttp pydantic is an aiohttp view to easily parse and validate request. You define using the function annotations what your methods for handling HTTP verbs expects and Aiohttp pydantic parses the HTTP request for you, validates the data, and injects that you want as parameters.

Features:

• Query string, request body, URL path and HTTP headers validation.
• Open API Specification generation.

How to install

$ pip install aiohttp_pydantic

Example:

from typing import Optional

from aiohttp import web
from aiohttp_pydantic import PydanticView
from pydantic import BaseModel

# Use pydantic BaseModel to validate request body
class ArticleModel(BaseModel):
    name: str
    nb_page: Optional[int]


# Create your PydanticView and add annotations.
class ArticleView(PydanticView):

    async def post(self, article: ArticleModel):
        return web.json_response({'name': article.name,
                                  'number_of_page': article.nb_page})

    async def get(self, with_comments: bool=False):
        return web.json_response({'with_comments': with_comments})


app = web.Application()
app.router.add_view('/article', ArticleView)
web.run_app(app)

$ curl -X GET http://127.0.0.1:8080/article?with_comments=a
[
  {
    "in": "query string",
    "loc": [
      "with_comments"
    ],
    "msg": "Input should be a valid boolean, unable to interpret input",
    "input": "a",
    "type": "bool_parsing"
  }
]

$ curl -X GET http://127.0.0.1:8080/article?with_comments=yes
{"with_comments": true}

$ curl -H "Content-Type: application/json" -X POST http://127.0.0.1:8080/article --data '{}'
[
  {
    "in": "body",
    "loc": [
      "name"
    ],
    "msg": "Field required",
    "input": {},
    "type": "missing"
  },
  {
    "in": "body",
    "loc": [
      "nb_page"
    ],
    "msg": "Field required",
    "input": {},
    "type": "missing"
  }
]

$ curl -H "Content-Type: application/json" -X POST http://127.0.0.1:8080/article --data '{"name": "toto", "nb_page": "3"}'
{"name": "toto", "number_of_page": 3}

Example using view function handler

from typing import Optional

from aiohttp import web
from aiohttp_pydantic.decorator import inject_params
from pydantic import BaseModel


# Use pydantic BaseModel to validate request body
class ArticleModel(BaseModel):
    name: str
    nb_page: Optional[int]


# Create your function decorated by 'inject_params' and add annotations.
@inject_params
async def post(article: ArticleModel):
    return web.json_response({'name': article.name,
                              'number_of_page': article.nb_page})


# If you need request
@inject_params.and_request
async def get(request, with_comments: bool = False):
    request.app["logger"]("OK")
    return web.json_response({'with_comments': with_comments})


app = web.Application()
app["logger"] = print
app.router.add_post('/article', post)
app.router.add_get('/article', get)
web.run_app(app)

API:

Inject Path Parameters

To declare a path parameter, you must declare your argument as a positional-only parameters:

Example:

class AccountView(PydanticView):
    async def get(self, customer_id: str, account_id: str, /):
        ...

app = web.Application()
app.router.add_get('/customers/{customer_id}/accounts/{account_id}', AccountView)

Inject Query String Parameters

To declare a query parameter, you must declare your argument as a simple argument:

class AccountView(PydanticView):
    async def get(self, customer_id: Optional[str] = None):
        ...

app = web.Application()
app.router.add_get('/customers', AccountView)

A query string parameter is generally optional and we do not want to force the user to set it in the URL. It's recommended to define a default value. It's possible to get a multiple value for the same parameter using the List type

from typing import List
from pydantic import Field

class AccountView(PydanticView):
    async def get(self, tags: List[str] = Field(default_factory=list)):
        ...

app = web.Application()
app.router.add_get('/customers', AccountView)

Inject Request Body

To declare a body parameter, you must declare your argument as a simple argument annotated with pydantic Model.

class Customer(BaseModel):
    first_name: str
    last_name: str

class CustomerView(PydanticView):
    async def post(self, customer: Customer):
        ...

app = web.Application()
app.router.add_view('/customers', CustomerView)

Inject HTTP headers

To declare a HTTP headers parameter, you must declare your argument as a keyword-only argument.

class CustomerView(PydanticView):
    async def get(self, *, authorization: str, expire_at: datetime):
        ...

app = web.Application()
app.router.add_view('/customers', CustomerView)

File Upload

You can receive files in addition to Pydantic data in your views. Here’s an example of how to use it: Usage Example

Suppose you want to create an API that accepts a book (with a title and a number of pages) as well as two files representing the pages of the book. Here’s how you can do it:

from aiohttp import web
from aiohttp_pydantic import PydanticView
from aiohttp_pydantic.uploaded_file import UploadedFile
from pydantic import BaseModel

class BookModel(BaseModel):
    title: str
    nb_page: int

class BookAndUploadFileView(PydanticView):
    async def post(self, book: BookModel, page_1: UploadedFile, page_2: UploadedFile):
        content_1 = (await page_1.read()).decode("utf-8")
        content_2 = (await page_2.read()).decode("utf-8")
        return web.json_response(
            {"book": book.model_dump(), "content_1": content_1, "content_2": content_2},
            status=201,
        )

Implementation Details

Files are represented by instances of UploadedFile, which wrap an aiohttp.BodyPartReader. UploadedFile exposes the read() and read_chunk() methods, allowing you to read the content of uploaded files asynchronously. You can use read() to get the complete content or read_chunk() to read chunks of data at a time.

Constraints to Consider

1 - Argument Order: If you use both Pydantic models and UploadedFile, you must always define BaseModel type arguments before UploadedFile type arguments. This ensures proper processing of the data.

2 - File Reading Order: UploadedFile instances must be read in the order they are declared in the method. Since files are not pre-loaded in memory or on disk, it is important to respect this order. If the reading order is not respected, a MultipartReadingError is raised.

Add route to generate Open Api Specification (OAS)

aiohttp_pydantic provides a sub-application to serve a route to generate Open Api Specification reading annotation in your PydanticView. Use aiohttp_pydantic.oas.setup() to add the sub-application

from aiohttp import web
from aiohttp_pydantic import oas


app = web.Application()
oas.setup(app)

By default, the route to display the Open Api Specification is /oas but you can change it using url_prefix parameter

oas.setup(app, url_prefix='/spec-api')

If you want generate the Open Api Specification from specific aiohttp sub-applications. on the same route, you must use apps_to_expose parameter.

from aiohttp import web
from aiohttp_pydantic import oas

app = web.Application()
sub_app_1 = web.Application()
sub_app_2 = web.Application()

oas.setup(app, apps_to_expose=[sub_app_1, sub_app_2])

You can change the title or the version of the generated open api specification using title_spec and version_spec parameters:

oas.setup(app, title_spec="My application", version_spec="1.2.3")

Add annotation to define response content

The module aiohttp_pydantic.oas.typing provides class to annotate a response content.

For example r200[List[Pet]] means the server responses with the status code 200 and the response content is a List of Pet where Pet will be defined using a pydantic.BaseModel

The docstring of methods will be parsed to fill the descriptions in the Open Api Specification.

from aiohttp_pydantic import PydanticView
from aiohttp_pydantic.oas.typing import r200, r201, r204, r404


class Pet(BaseModel):
    id: int
    name: str


class Error(BaseModel):
    error: str


class PetCollectionView(PydanticView):
    async def get(self) -> r200[List[Pet]]:
        """
        Find all pets

        Tags: pet
        """
        pets = self.request.app["model"].list_pets()
        return web.json_response([pet.dict() for pet in pets])

    async def post(self, pet: Pet) -> r201[Pet]:
        """
        Add a new pet to the store

        Tags: pet
        Status Codes:
            201: The pet is created
        """
        self.request.app["model"].add_pet(pet)
        return web.json_response(pet.dict())


class PetItemView(PydanticView):
    async def get(self, id: int, /) -> Union[r200[Pet], r404[Error]]:
        """
        Find a pet by ID

        Tags: pet
        Status Codes:
            200: Successful operation
            404: Pet not found
        """
        pet = self.request.app["model"].find_pet(id)
        return web.json_response(pet.dict())

    async def put(self, id: int, /, pet: Pet) -> r200[Pet]:
        """
        Update an existing pet

        Tags: pet
        Status Codes:
            200: successful operation
        """
        self.request.app["model"].update_pet(id, pet)
        return web.json_response(pet.dict())

    async def delete(self, id: int, /) -> r204:
        self.request.app["model"].remove_pet(id)
        return web.Response(status=204)

Group parameters

If your method has lot of parameters you can group them together inside one or several Groups.

from aiohttp_pydantic.injectors import Group

class Pagination(Group):
    page_num: int = 1
    page_size: int = 15


class ArticleView(PydanticView):

    async def get(self, page: Pagination):
        articles = Article.get(page.page_num, page.page_size)
        ...

The parameters page_num and page_size are expected in the query string, and set inside a Pagination object passed as page parameter.

The code above is equivalent to:

class ArticleView(PydanticView):

    async def get(self, page_num: int = 1, page_size: int = 15):
        articles = Article.get(page_num, page_size)
        ...

You can add methods or properties to your Group.

class Pagination(Group):
    page_num: int = 1
    page_size: int = 15

    @property
    def num(self):
        return self.page_num

    @property
    def size(self):
        return self.page_size

    def slice(self):
        return slice(self.num, self.size)


class ArticleView(PydanticView):

    async def get(self, page: Pagination):
        articles = Article.get(page.num, page.size)
        ...

Custom Validation error

You can redefine the on_validation_error hook in your PydanticView

class PetView(PydanticView):

    async def on_validation_error(self,
                                  exception: ValidationError,
                                  context: str):
        errors = exception.errors()
        for error in errors:
            error["in"] = context  # context is "body", "headers", "path" or "query string"
            error["custom"] = "your custom field ..."
        return json_response(data=errors, status=400)

If you use function based view:

async def custom_error(exception: ValidationError,
                       context: str):
    errors = exception.errors()
    for error in errors:
        error["in"] = context  # context is "body", "headers", "path" or "query string"
        error["custom"] = "your custom field ..."
    return json_response(data=errors, status=400)


@inject_params(on_validation_error=custom_error)
async def get(with_comments: bool = False):
    ...

@inject_params.and_request(on_validation_error=custom_error)
async def get(request, with_comments: bool = False):
    ...

A tip to use the same error handling on each view

inject_params = inject_params(on_validation_error=custom_error)


@inject_params
async def post(article: ArticleModel):
    return web.json_response({'name': article.name,
                              'number_of_page': article.nb_page})


@inject_params.and_request
async def get(request, with_comments: bool = False):
    return web.json_response({'with_comments': with_comments})

Add security to the endpoints

aiohttp_pydantic provides a basic way to add security to the endpoints. You can define the security on the setup level using the security parameter and then mark view methods that will require this security schema.

from aiohttp import web
from aiohttp_pydantic import oas


app = web.Application()
oas.setup(app, security={"APIKeyHeader": {"type": "apiKey", "in": "header", "name": "Authorization"}})

And then mark the view method with the security descriptor

from aiohttp_pydantic import PydanticView
from aiohttp_pydantic.oas.typing import r200, r201, r204, r404


class Pet(BaseModel):
    id: int
    name: str


class Error(BaseModel):
    error: str


class PetCollectionView(PydanticView):
    async def get(self) -> r200[List[Pet]]:
        """
        Find all pets

        Security: APIKeyHeader
        Tags: pet
        """
        pets = self.request.app["model"].list_pets()
        return web.json_response([pet.dict() for pet in pets])

    async def post(self, pet: Pet) -> r201[Pet]:
        """
        Add a new pet to the store

        Tags: pet
        Status Codes:
            201: The pet is created
        """
        self.request.app["model"].add_pet(pet)
        return web.json_response(pet.dict())

Demo

Have a look at demo for a complete example

git clone https://github.com/Maillol/aiohttp-pydantic.git
cd aiohttp-pydantic
pip install .
python -m demo

Go to

You can generate the OAS in a json or yaml file using the aiohttp_pydantic.oas command:

python -m aiohttp_pydantic.oas demo.main

$ python3 -m aiohttp_pydantic.oas  --help
usage: __main__.py [-h] [-b FILE] [-o FILE] [-f FORMAT] [APP [APP ...]]

Generate Open API Specification

positional arguments:
  APP                   The name of the module containing the asyncio.web.Application. By default the variable named
                        'app' is loaded but you can define an other variable name ending the name of module with :
                        characters and the name of variable. Example: my_package.my_module:my_app If your
                        asyncio.web.Application is returned by a function, you can use the syntax:
                        my_package.my_module:my_app()

optional arguments:
  -h, --help            show this help message and exit
  -b FILE, --base-oas-file FILE
                        A file that will be used as base to generate OAS
  -o FILE, --output FILE
                        File to write the output
  -f FORMAT, --format FORMAT
                        The output format, can be 'json' or 'yaml' (default is json)