Typed JSON dataclasses deserialisation for Python 3.9+
pip install jsondataclasses
from datetime import date
from typing import Literal, Optional
from jsondataclasses import jsondataclass, jsonfield
def parse_date(date_string: str) -> date:
return date.fromisoformat(date_string.replace(".", "-"))
@jsondataclass
class Car:
make: Literal["Ford", "Renault", "Volkswagen"] = jsonfield("carMake")
model: str = jsonfield("model")
manufactured_at: date = jsonfield("dateOfManufacture", parse_date)
num_of_wheels: Optional[int] = jsonfield("numberOfWheels", default_value=4)
car = Car({
"carMake": "Ford",
"model": "Focus",
"dateOfManufacture": "2018.03.14"
})
print(car) # Car(make='Ford', model='Focus', manufactured_at=datetime.date(2018, 3, 14), num_of_wheels=4)
Class field types can be any primitive type (eg. str
, int
, datetime
),
a variadic generic (eg. list[str]
, Optional[int]
, Literal["hello", "world"]
),
or even another jsondataclass. The default_value
argument of jsonfield
will be used
if the specified key is not found in the dictionary and the type of the class field is not Optional[...]
.
In this case, the value of default_value
will be passed to the supplied parser function.