Allows you to convert pydantic models for fastapi param models - query, form, header, cookie, body, etc.
pip install pyfa_converter
fromdatetimeimportdatetimefromtypingimportOptionalfromfastapiimportFastAPI, UploadFile, File, FormfrompydanticimportBaseModel, Fieldfrompyfa_converterimportFormDepends, PyFaDependsapp=FastAPI()
classPostContractBodySchema(BaseModel):
title: str=Field(..., description="Description title")
date: Optional[datetime] =Field(
None, description="Example: 2021-12-14T09:56:31.056Z"
)
@app.post("/form-data-body")asyncdefexample_foo_body_handler(
data1: PostContractBodySchema=PyFaDepends(model=PostContractBodySchema, _type=Form),
document: UploadFile=File(...),
):
return {"title": data.title, "date": data.date, "file_name": document.filename}frompyfa_converterimportPyFaDepends, FormDepends, QueryDependsfromfastapiimportHeader, Form
...
asyncdeffoo(data: MyCustomModel=PyFaDepends(MyCustomModel, _type=Header)): ...
asyncdeffoo(data: MyCustomModel=PyFaDepends(MyCustomModel, _type=Form)): ...
asyncdeffoo(data: MyCustomModel=FormDepends(MyCustomModel)): ...
asyncdeffoo(data: MyCustomModel=QueryDepends(MyCustomModel)): ...If you want to accept a file on an endpoint, then the content-type for that endpoint changes from application/json to www-form-data.
FastAPI does not know how to override the pydantic schema so that parameters are passed as form. Even if you do
foo: CustomPydanticModel = Depends()
all model attributes will be passed as query, but we want them to become body, that's what this library exists for.
But, if we accept a lot of fields, then the function becomes very large (the number of attributes for the endpoint increases and it does not look very good).
Thanks to this library, it is possible to force the conversion of Field fields into fields of FastAPI Form with full preservation of all attributes (alias, gt, te, description, title, example and more...)
