Cannot use a pydantic dataclass nested in a BaseModel as a response_model

[ypsah]

First Check

• I added a very descriptive title to this issue.
• I used the GitHub search to find a similar issue and didn't find it.
• I searched the FastAPI documentation, with the integrated search.
• I already searched in Google "How to X in FastAPI" and didn't find any information.
• I already read and followed all the tutorial in the docs and didn't find an answer.
• I already checked if it is not related to FastAPI but to Pydantic.
• I already checked if it is not related to FastAPI but to Swagger UI.
• I already checked if it is not related to FastAPI but to ReDoc.

Commit to Help

• I commit to help with one of those options 👆

Example Code

from fastapi import FastAPI
from pydantic import BaseModel
from pydantic.dataclasses import dataclass


app = FastAPI()


@dataclass
class Item:
    name: str


class Box(BaseModel):
    item: Item


@app.get("/item", response_model=Item)
async def item() -> Item:
    return Item(name="test")


@app.get("/box", response_model=Box)
async def box() -> Box:
    return Box(item=await item())

Description

The item endpoint works fine, but the box one fails with a ValidationError:

$ uvicorn main:app
INFO:     Started server process [13673]
INFO:     Waiting for application startup.
INFO:     Application startup complete.
INFO:     Uvicorn running on http://127.0.0.1:8000 (Press CTRL+C to quit)
INFO:     127.0.0.1:46020 - "GET /docs HTTP/1.1" 200 OK
INFO:     127.0.0.1:46020 - "GET /openapi.json HTTP/1.1" 200 OK
INFO:     127.0.0.1:46020 - "GET /item HTTP/1.1" 200 OK
INFO:     127.0.0.1:46020 - "GET /box HTTP/1.1" 500 Internal Server Error
ERROR:    Exception in ASGI application
Traceback (most recent call last):
  File "/home/guest/.local/share/virtualenvs/tmp.KhplXQz0Am-XrtSmM4b/lib/python3.10/site-packages/uvicorn/protocols/http/h11_impl.py", line 373, in run_asgi
    result = await app(self.scope, self.receive, self.send)
  File "/home/guest/.local/share/virtualenvs/tmp.KhplXQz0Am-XrtSmM4b/lib/python3.10/site-packages/uvicorn/middleware/proxy_headers.py", line 75, in __call__
    return await self.app(scope, receive, send)
  File "/home/guest/.local/share/virtualenvs/tmp.KhplXQz0Am-XrtSmM4b/lib/python3.10/site-packages/fastapi/applications.py", line 208, in __call__
    await super().__call__(scope, receive, send)
  File "/home/guest/.local/share/virtualenvs/tmp.KhplXQz0Am-XrtSmM4b/lib/python3.10/site-packages/starlette/applications.py", line 112, in __call__
    await self.middleware_stack(scope, receive, send)
  File "/home/guest/.local/share/virtualenvs/tmp.KhplXQz0Am-XrtSmM4b/lib/python3.10/site-packages/starlette/middleware/errors.py", line 181, in __call__
    raise exc
  File "/home/guest/.local/share/virtualenvs/tmp.KhplXQz0Am-XrtSmM4b/lib/python3.10/site-packages/starlette/middleware/errors.py", line 159, in __call__
    await self.app(scope, receive, _send)
  File "/home/guest/.local/share/virtualenvs/tmp.KhplXQz0Am-XrtSmM4b/lib/python3.10/site-packages/starlette/exceptions.py", line 82, in __call__
    raise exc
  File "/home/guest/.local/share/virtualenvs/tmp.KhplXQz0Am-XrtSmM4b/lib/python3.10/site-packages/starlette/exceptions.py", line 71, in __call__
    await self.app(scope, receive, sender)
  File "/home/guest/.local/share/virtualenvs/tmp.KhplXQz0Am-XrtSmM4b/lib/python3.10/site-packages/starlette/routing.py", line 656, in __call__
    await route.handle(scope, receive, send)
  File "/home/guest/.local/share/virtualenvs/tmp.KhplXQz0Am-XrtSmM4b/lib/python3.10/site-packages/starlette/routing.py", line 259, in handle
    await self.app(scope, receive, send)
  File "/home/guest/.local/share/virtualenvs/tmp.KhplXQz0Am-XrtSmM4b/lib/python3.10/site-packages/starlette/routing.py", line 61, in app
    response = await func(request)
  File "/home/guest/.local/share/virtualenvs/tmp.KhplXQz0Am-XrtSmM4b/lib/python3.10/site-packages/fastapi/routing.py", line 234, in app
    response_data = await serialize_response(
  File "/home/guest/.local/share/virtualenvs/tmp.KhplXQz0Am-XrtSmM4b/lib/python3.10/site-packages/fastapi/routing.py", line 137, in serialize_response
    raise ValidationError(errors, field.type_)
pydantic.error_wrappers.ValidationError: 1 validation error for Box
response -> item
  value is not a valid dict (type=type_error.dict)

Operating System

Linux

Operating System Details

No response

FastAPI Version

0.71.0

Python Version

3.10.1

Additional Context

No response

[harunyasar]

As far as I understand, FastAPI doesn't let us to mix BaseModel with either pydantic dataclass or standard library dataclass. You can only mix same type of models.

import dataclasses

from pydantic.dataclasses import dataclass


@dataclass  # or @dataclasses.dataclass
class Item:
    name: str

@dataclass  # or @dataclasses.dataclass
class Box:
    item: Item

@app.get("/item", response_model=Item)
async def item() -> Item:
    return Item(name="test")


@app.get("/box", response_model=Box)
async def box() -> Box:
    return Box(item=await item())

[ypsah]

This isn't what fastapi's documentation implies:

You can also combine dataclasses with other Pydantic models, inherit from them, include them in your own models, etc.

ref: https://fastapi.tiangolo.com/advanced/dataclasses/?h=dataclasses#learn-more

pydantic's documentation gives an explicit example mixing the two here.

Also creating an instance of Box manually works without any hiccups, and I can even validate it myself:

>>> Box.validate(Box(item=Item(name="test")))
Box(item=Item(name="test")))

What makes you think mixing the two is not supported?

[harunyasar]

Neither pydantic nor FastAPI documentation made me think that mixing different types of models is not supported. I have only tried to mix BaseModel with pydantic dataclass and BaseModel with stdlib dataclass. Each time I tried I got the same error unless I returned a dict object explicitly Box(item=await item()).dict(). That's why I'm saying, apparently, FastAPI (not the pydantic) doesn't let us mix different types. I haven't checked the links you pointed out above but will do it later on.

I have checked the documentation but I couldn't see anything that says you could mix a BaseModel and a dataclass. I guess the intention was there to say that you could either use BaseModel or dataclass as an request or response model. I have checked the source code as well.

fastapi/fastapi/routing.py

Lines 75 to 80 in 672c55a

	return res.dict(
	by_alias=True,
	exclude_unset=exclude_unset,
	exclude_defaults=exclude_defaults,
	exclude_none=exclude_none,
	)

Since Box is a BaseModel, it immediately returns a dict object {"item": Item(name="test")} and it is not a valid dict.

If you call as I showed you above your response will be a dict object and dataclass.asdict() method of the dataclass will be called.

fastapi/fastapi/routing.py

Lines 101 to 102 in 672c55a

	elif dataclasses.is_dataclass(res):
	return dataclasses.asdict(res)

Eventually, this will also be a dict object that can be validated and accepted by the Box response object.

I might be wrong. I hope there will be someone else that can have a better explanation.

[ypsah]

Thanks for the pointers in the codebase! 👍

I am not sure this is the whole story, but it is definitely a step forward.
I think it is enough for me to carry on and send a PR to either pydantic or fastapi, maybe both.

---

I have checked the documentation but I couldn't see anything that says you could mix a BaseModel and a dataclass.

Well, there is the part that I posted in my first response that says:

You can also combine dataclasses with other Pydantic models, inherit from them, include them in your own models, etc.

With a link to that sentence.
Did you check it out?

---

For the record, editing a post does not trigger any notification.
I think it would have been better in this case to add a new comment.

[harunyasar]

Yes, I did. Like I explained above, those were my findings and I believe _prepare_response_content might need an improvement. If you are going to create a PR that can fix the issue it would be great and I would be glad to see. Apart from that, until having seen a better explanation I don’t have much to say.

[databasedav]

this is a bug pydantic/pydantic#3764