FastAPI: Using multiple dependencies (functions), each with pydantic request bodies

Question

With FastAPI I am trying to inject two dependent functions, each of which have their own pydantic request bodies and form part of separate endpoints into path operation of another endpoint. Please see the example below just for illustration:

from fastapi import FastAPI, Depends
from pydantic import BaseModel

app = FastAPI()

class RequestBodyChild1(BaseModel):
    item1: float
    item2: float

class RequestBody1(BaseModel):
    parent1: RequestBodyChild1

@app.post("/function_1")
def function_1(request: RequestBody1):
    return request

class RequestBodyChild2(BaseModel):
    itemA: float
    itemB: float

class RequestBody2(BaseModel):
    parent2: RequestBodyChild2


@app.post("/function_2")
def function_2(request: RequestBody2):
    return request


@app.post("/function_3")
def function_3(request_1: dict = Depends(function_1), request_2: dict = Depends(function_2)):
    merged_dict = dict(request_1.items() + request_2.items())
    return merged_dict

The issue is that for the function_3 endpoint OpenAPI/Swagger only shows the request body/example schema for request_1: dict = Depends(function_1):

{
  "parent1": {
    "item1": 0,
    "item2": 0
  }
}

How can I make sure that both request bodies are included? Along the lines of:

{
  "parent1": {
    "item1": 0,
    "item2": 0
  },
  "parent2": {
    "itemA": 0,
    "itemB": 0
  }
}

MatsLindh · Accepted Answer · 2021-10-29 19:02:59Z

1

The reason for the result you're seeing as that you're using the same parameter name in both functions - when you're using Depends in that way, the parameter name in the function is used (the function being registered as a view isn't relevant; only that it is a function); since it's the same in both cases, you only get a single argument name filled. You can see this change if you rename the parameter in function_2 to request2:

{
  "request": {
    "parent1": {
      "item1": 0,
      "item2": 0
    }
  },
  "request2": {
    "parent2": {
      "itemA": 0,
      "itemB": 0
    }
  }
}

To achieve the request format you've described I think you'd have to define the parent1: RequestBodyChild1, parent2: RequestBodyChild2 directly on function_3, and then call the dependent functions manually.

def function_3(parent1: RequestBodyChild1, parent2: RequestBodyChild2):
  ...

This resolves to:

{
  "parent1": {
    "item1": 0,
    "item2": 0
  },
  "parent2": {
    "itemA": 0,
    "itemB": 0
  }
}

answered Oct 29, 2021 at 19:02

MatsLindh

53.5k5 gold badges70 silver badges97 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

Svencken Over a year ago

Many thanks for this advice, I wasn't aware the identical parameter names in function_1 and function_2 was the underlying cause of this. Good to know. It's a pity that calling two dependencies doesn't result in the request body I was looking for. A single dependency as a single parameter in function_3 does result in the omission of the "parent1" level. I've gone with the manual solution in the end.

Collectives™ on Stack Overflow

FastAPI: Using multiple dependencies (functions), each with pydantic request bodies

1 Answer 1

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related