Multipart Form Data

Reading the request body as multipart/form-data.

Description

Form Data (MIME-type: multipart/form-data) is one of the most commonly used content types for sending binary data to a server.

The multipart format means that data is sent to the server in separate parts. Each part can have its own content type, filename, and data. Data is separated using a boundary string.

from pydantic import BaseModel, ConfigDict
from rapidy.http import post, Body, ContentType, FileField

class UserData(BaseModel):
    model_config = ConfigDict(arbitrary_types_allowed=True)

    username: str
    password: str
    image: FileField

@post('/')
async def handler(
    user_data: UserData = Body(content_type=ContentType.m_part_form_data),
    # or
    user_data: UserData = Body(content_type='multipart/form-data'),
) -> ...:

Data Example

POST /  HTTP/1.1
Host: localhost:8080
Content-Type: multipart/form-data; boundary=---WD9146A
Content-Length: ...

---WD9146A
Content-Disposition: form-data; name="username"

User
---WD9146A
Content-Disposition: form-data; name="password"

myAwesomePass
---WD9146A
Content-Disposition: form-data; name="image"; filename="image.png";
Content-Type: image/png

<... binary data ...>
---WD9146A

Sending with curl

curl -X POST \
-H "Content-Type: multipart/form-data" \
-F username=User \
-F password=myAwesomePass \
http://127.0.0.1:8080

---

Extraction Without Validation

Disabling validation is not recommended.

If validation is disabled, the parameter will contain the base aiohttp structure:

• Body(content_type=ContentType.m_part_form_data) → MultiDictProxy[Union[str, bytes, FileField]]

Ways to Disable Validation

Explicit Disabling

from pydantic import BaseModel
from rapidy.http import post, Body, ContentType

class BodyData(BaseModel):
    ...

@post('/')
async def handler(
    data: BodyData = Body(validate=False, content_type=ContentType.m_part_form_data),
) -> ...:

Using Any

@post('/')
async def handler(
    data: Any = Body(content_type=ContentType.m_part_form_data),
) -> ...:

No Type Annotation

If no type is specified, Any will be set by default.

@post('/')
async def handler(
    data=Body(content_type=ContentType.m_part_form_data),
) -> ...:

---

Default Values

If an HTTP request does not contain a body, the parameter will receive the specified default value (if set).

Usage Examples

Default Value Specified

from pydantic import BaseModel
from rapidy.http import post, Body, ContentType

class BodyData(BaseModel):
    ...

@post('/')
async def handler(
    data: BodyData = Body('some_data', content_type=ContentType.m_part_form_data),
    # or
    data: BodyData = Body(default_factory=lambda: 'some_data', content_type=ContentType.m_part_form_data),
) -> ...:

Optional Request Body

from pydantic import BaseModel
from rapidy.http import post, Body, ContentType

class BodyData(BaseModel):
    ...

@post('/')
async def handler(
    data: BodyData | None = Body(content_type=ContentType.m_part_form_data),
    # or
    data: Optional[BodyData] = Body(content_type=ContentType.m_part_form_data),
    # or
    data: Union[BodyData, None] = Body(content_type=ContentType.m_part_form_data),
) -> ...:

---

Extracting Raw Data

Rapidy uses the post method of the Request object to obtain data and passes it to Pydantic for validation.

How data extraction works in Rapidy

async def extract_post_data(request: Request) -> Optional[MultiDictProxy[Union[str, bytes, FileField]]]:
    if not request.body_exists:
        return None

    return await request.post()

Rapidy uses built-in aiohttp mechanisms for data extraction.

More details about the aiohttp.Request object and methods for extracting data from it can be found here.

x-www-form-urlencoded and multipart/form-data are processed the same way.

Both of these content types are extracted using the post method of the Request object. This is a feature of aiohttp.

If a parameter is annotated as bytes or StreamReader, data is extracted differently.

More details about the StreamReader object can be found here.

bytes

from rapidy.http import post, Body, ContentType

@post('/')
async def handler(
    user_data: bytes = Body(content_type=ContentType.m_part_form_data),
    # also you can use pydantic validation
    user_data: bytes = Body(content_type=ContentType.m_part_form_data, min_length=1),
) -> ...:

Rapidy Internal Code

async def extract_body_bytes(request: Request) -> Optional[bytes]:
    if not request.body_exists:
        return None

    return await request.read()

StreamReader

from rapidy import StreamReader
from rapidy.http import post, Body, ContentType

@post('/')
async def handler(
    user_data: StreamReader = Body(content_type=ContentType.m_part_form_data),
) -> ...:

Rapidy Internal Code

async def extract_body_stream(request: Request) -> Optional[StreamReader]:
    if not request.body_exists:
        return None

    return request.content

Validation with Pydantic is not supported for StreamReader.

Default values cannot be set for StreamReader.

If you attempt to set a default value for Body with a StreamReader annotation using default or default_factory, a ParameterCannotUseDefaultError will be raised.

from rapidy import StreamReader
from rapidy.http import post, Body, ContentType

@post('/')
async def handler(
    user_data: StreamReader = Body(default='SomeDefault', content_type=ContentType.m_part_form_data),
) -> ...:

------------------------------
Handler attribute with Type `Body` cannot have a default value.

Handler path: `<full_path>/main.py`
Handler name: `handler`
Attribute name: `data`
------------------------------