OpenAPI generation fails with dates with Annotated constraint

[jereflit]

First Check

• I added a very descriptive title here.
• I used the GitHub search to find a similar question 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 datetime import date
from typing import Annotated

import annotated_types
from fastapi import FastAPI
from pydantic import BaseModel

app = FastAPI()


class DataInput(BaseModel):
    birthdate: Annotated[
        date,
        annotated_types.Gt(date(1900, 1, 1)),
        annotated_types.Lt(date(2000, 1, 1)),
    ]


@app.post("/")
def post_root(
    data_input: DataInput,
):
    return {"birthdate": data_input.birthdate}

EDIT: the birthdate field must be a date field of a pydantic model, as is it used in a POST endpoint

Description

A pydantic ValidationError arises when FastApi calls the openapi utils that generates an instance of OpenAPI

fastapi/fastapi/openapi/utils.py

Line 530 in a9819df

return jsonable_encoder(OpenAPI(**output), by_alias=True, exclude_none=True) # type: ignore

  ...
  File "/home/jereflit/bug/.venv/lib/python3.11/site-packages/fastapi/applications.py", line 1009, in openapi
    return JSONResponse(self.openapi())
                        ^^^^^^^^^^^^^^
  File "/home/jereflit/bug/.venv/lib/python3.11/site-packages/fastapi/applications.py", line 981, in openapi
    self.openapi_schema = get_openapi(
                          ^^^^^^^^^^^^
  File "/home/jereflit/bug/.venv/lib/python3.11/site-packages/fastapi/openapi/utils.py", line 530, in get_openapi
    return jsonable_encoder(OpenAPI(**output), by_alias=True, exclude_none=True)  # type: ignore
                            ^^^^^^^^^^^^^^^^^
  File "/home/jereflit/bug/.venv/lib/python3.11/site-packages/pydantic/main.py", line 176, in __init__
    self.__pydantic_validator__.validate_python(data, self_instance=self)

pydantic_core._pydantic_core.ValidationError: 4 validation errors for OpenAPI
components.schemas.PricerInput.Schema.properties.driver_birth_date.Schema.exclusiveMaximum
  Input should be a valid number [type=float_type, input_value=datetime.date(2006, 6, 5), input_type=date]
    For further information visit https://errors.pydantic.dev/2.7/v/float_type
components.schemas.PricerInput.Schema.properties.driver_birth_date.Schema.exclusiveMinimum
  Input should be a valid number [type=float_type, input_value=datetime.date(2005, 6, 6), input_type=date]
    For further information visit https://errors.pydantic.dev/2.7/v/float_type
components.schemas.PricerInput.Schema.properties.driver_birth_date.bool
  Input should be a valid boolean [type=bool_type, input_value={'description': 'Date of ...Date', 'type': 'string'}, input_type=dict]
    For further information visit https://errors.pydantic.dev/2.7/v/bool_type
components.schemas.PricerInput.Reference.$ref
  Field required [type=missing, input_value={'properties': {'bonus_ma...nput', 'type': 'object'}, input_type=dict]
    For further information visit https://errors.pydantic.dev/2.7/v/missing

Note that the concerned fields are indeed defined as optional floats :

fastapi/fastapi/openapi/models.py

Lines 164 to 167 in a9819df

	maximum: Optional[float] = None
	exclusiveMaximum: Optional[float] = None
	minimum: Optional[float] = None
	exclusiveMinimum: Optional[float] = None

because OpenAPI is not supposed to support values restriction for dates, which is confirmed here : OAI/OpenAPI-Specification#2827

However, FastAPI raises a ValidationError and makes /openapi.json fails, and therefore Swagger /docs and Redoc /redoc fail too, whereas the annotated type is correctly defined.

In the OpenAPi discussion linked above, someone suggests to use custom keywords minimumDate and maximumDate.
👉 Is it something that I can explicitly do on my side by writing my own get_openapi function ?

(A suggested workaround would be to handle the validation myself by adding a pydantic's @field_validator to my DataInput class. It is fine, except that neither OpenAPI json nor the type checker will reflect those constraints on input data.)

Operating System

Linux

Operating System Details

No response

FastAPI Version

0.111.0

Pydantic Version

2.7.3

Python Version

Python 3.11.5

Additional Context

No response

[JavierSanchezCastro]

With datetime you can convert it to timestamp and set the gt and lt in the Query param

from datetime import date, datetime
from typing import Annotated

from fastapi import FastAPI, Query

app = FastAPI()


@app.get("/")
def read_root(
    data_input: Annotated[date, Query(gt=datetime(1900, 1, 1).timestamp(), lt=datetime(2000, 1, 1).timestamp())]
):
    return {"birthdate": data_input}

[jereflit]

Thanks for your help !

It does work like in your example (even with a regular date):

from datetime import date
from typing import Annotated

import annotated_types
from fastapi import FastAPI

app = FastAPI()


@app.get("/")
def read_root(
    birthdate: Annotated[date, annotated_types.Gt(date(1900, 1, 1)), annotated_types.Lt(date(2000, 1, 1))],
):
    return {"birthdate": birthdate}

but this is because birthdate is defined here as a Query param, which seems to be supported by the OpenAPI generator :

...
"parameters": [
    {
        "name": "birthdate",
        "in": "query",
        "required": true,
        "schema": {
            "type": "string",
            "format": "date",
            "exclusiveMaximum": "2000-01-01",
            "exclusiveMinimum": "1900-01-01",
            "title": "Birthdate"
        }
    }
],
...

However, in my case, birthdate is a field of a Body param defined in a pydantic model, and needs to remain that way.

[JavierSanchezCastro]

Well if it's a Body param you need to use POST instead of GET. Taking that in mind, you can rewrite the code as

from datetime import date, datetime
from typing import Annotated
from pydantic import BaseModel, Field
from fastapi import FastAPI

app = FastAPI()

class DataInput(BaseModel):
    birthdate: Annotated[date, Field(gt=datetime(1900, 1, 1).timestamp(), lt=datetime(2000, 1, 1).timestamp())]

@app.post("/")
def read_root(
    data_input: DataInput
):
    return {"birthdate": data_input.birthdate}

[jereflit]

My bad, it should have been a POST from the beginning indeed.

However, what you suggest does not work, because datetime is not compatible with date :

pydantic_core._pydantic_core.SchemaError: Invalid Schema:
model.schema.model-fields.fields.birthdate.schema.date.lt
  Datetimes provided to dates should have zero time - e.g. be exact dates [type=date_from_datetime_inexact, input_value=946681200.0, input_type=float]
    For further information visit https://errors.pydantic.dev/2.7/v/date_from_datetime_inexact
model.schema.model-fields.fields.birthdate.schema.date.gt
  Datetimes provided to dates should have zero time - e.g. be exact dates [type=date_from_datetime_inexact, input_value=-2208989361.0, input_type=float]
    For further information visit https://errors.pydantic.dev/2.7/v/date_from_datetime_inexact

unless as I accept birthdate to be given as datetimes... but this is not my use case.

[JavierSanchezCastro]

I've tried with your versions and it's working well

pydantic==2.7.3
fastapi==0.111.0

The type for birthdate is date, but just for the verification I used datetime

[jereflit]

I don't understand. I cannot even launch the api with your code :

 ~/bug   uvicorn main:app --reload

INFO:     Will watch for changes in these directories: ['/home/jereflit/bug']
INFO:     Uvicorn running on http://127.0.0.1:8000 (Press CTRL+C to quit)
INFO:     Started reloader process [175641] using WatchFiles
Process SpawnProcess-1:
Traceback (most recent call last):
  File "/home/jereflit/.pyenv/versions/3.11.5/lib/python3.11/multiprocessing/process.py", line 314, in _bootstrap
    self.run()
  
  [.....skipped.....]

  File "/home/jereflit/bug/main.py", line 8, in <module>
    class DataInput(BaseModel):
  File "/home/jereflit/bug/.venv/lib/python3.11/site-packages/pydantic/_internal/_model_construction.py", line 202, in __new__
    complete_model_class(
  File "/home/jereflit/bug/.venv/lib/python3.11/site-packages/pydantic/_internal/_model_construction.py", line 549, in complete_model_class
    schema = gen_schema.clean_schema(schema)
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/jereflit/bug/.venv/lib/python3.11/site-packages/pydantic/_internal/_generate_schema.py", line 442, in clean_schema
    schema = validate_core_schema(schema)
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/jereflit/bug/.venv/lib/python3.11/site-packages/pydantic/_internal/_core_utils.py", line 568, in validate_core_schema
    return _validate_core_schema(schema)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
pydantic_core._pydantic_core.SchemaError: Invalid Schema:
model.schema.model-fields.fields.birthdate.schema.date.lt
  Datetimes provided to dates should have zero time - e.g. be exact dates [type=date_from_datetime_inexact, input_value=946681200.0, input_type=float]
    For further information visit https://errors.pydantic.dev/2.7/v/date_from_datetime_inexact
model.schema.model-fields.fields.birthdate.schema.date.gt
  Datetimes provided to dates should have zero time - e.g. be exact dates [type=date_from_datetime_inexact, input_value=-2208989361.0, input_type=float]
    For further information visit https://errors.pydantic.dev/2.7/v/date_from_datetime_inexact

and yet I just copied/pasted it

[JavierSanchezCastro]

I can't reproduce your error with the versions you provide