Integration with FastAPI

PonyORM can be used with FastAPI, it integrates with Pydantic and can be used in an async environment, as long as you follow a few rules.


For your first steps, simply follow Getting Started with Pony to create the database object and define your models. You may want to make use of Pydantic’s inbuilt BaseSettings Model to manage your connection secrets. The specific parameters can be found in Connecting to the Database.

import fastapi
from src.config import Settings
from src.models import db

settings = Settings()
api = fastapi.FastAPI()

Models and Schemas

To make full use of FastAPI, Pony’s models have to work with Pydantic’s schemas. Unfortunately, this requires somewhat duplicate model definitions. We’re following FastAPI’s example here.

Here are the Pony models from Getting Started with Pony:

class Person(db.Entity):
    name = Required(str)
    age = Required(int)
    cars = Set('Car')

class Car(db.Entity):
    make = Required(str)
    model = Required(str)
    owner = Required(Person)

We want to expose Person in FastAPI like this:

async def read_all_persons():
    with db_session:
       persons =
       result = [PersonInDB.from_orm(p) for p in persons]
    return result

async def read_single_person(pid: int):
    with db_session:
        person = Person[pid]
        result = PersonInDB.from_orm(person)
    return result

For this, we can write the following Pydantic schemas:

class CarOut(BaseModel):
    make: str
    model: str

class PersonInDB(BaseModel):
    name: str
    age: int
    cars: List[CarOut]

    @validator('cars', pre=True, allow_reuse=True)
    def pony_set_to_list(cls, values):
        return [v.to_dict() for v in values]

    class Config:
        orm_mode = True

Without preparation, Pony’s related objects aren’t interpreted correctly. Use a validator to make them accessible for Pydantic.

Async and db_session

Pony was not developed for async usage, and it may be tricky to use it correctly in an async environment. However, if only part of your application is async, you will be fine if you stick to these two rules:

You cannot use the @db_session decorator on async functions, use a context manager for wrapping lines of code that really work with the database. Secondly, do not call async functions inside a db_session, only normal function calls are possible. Use little shortliving sessions and don’t interrupt them with async:

async def func():
    with db_session:
        thing = database.MyEntity()
    await async_func(ABC, thing)