I'm making an app that would allow you to either log in or create a restaurant. As a restaurant owner, you can add/remove/edit refrigerators. My end goal is that I'd have a list of Restaurants that I'd write to a JSON file, and anytime I rerun the app, I can pull that data in and simulate "being a restaurant owner" and edit the fridges for the chosen restaurant.
I essentially want this:
data = {
restaurants: [
{
restaurant: {
name: "Peppy",
pw: "123",
fridges: [
{
fridge: {
owner: restaurant.name,
contents: []
}
}
]
}
}
]
}
I have the following two classes(showing relevant methods):
class Restaurant:
def __init__(self, owner, password):
self.__password = password
self.owner = owner
self.__owned_fridges = [] # list of owned fridges
def add_fridge(self):
nickname = input("What would you like to name the fridge: ")
self.__owned_fridges.append(fr(self.owner, nickname))
print("Fridge added!")
class Fridge:
def __init__(self, owner, nickname):
self.nickname = nickname
self.owner = owner
self.__authorized_users = [owner]
self.__contents = []
def add_to_fridge(self):
if len(self.__contents) == 5:
print("Your fridge is full!")
else:
item = input("What would you like to add : ")
self.__contents.append(item)
My issue is in serializing this for JSON. I have found that the following works to serialize the restaurant object to JSON, but not the nested fridge objects :
data = {
'restaurants': []
}
# Testing code
test = res("Jac", "350b534")
test.add_fridge()
test.add_fridge()
data['restaurants'].append(json.dumps(test.__dict__))
I'm relatively new to python, and I come from a js background, so I'm still getting familiar with the syntax. My question is, how do I serialize the inner list of fridges?
CodePudding user response:
Firstly, I'd suggest checking out dataclasses as they will simplify the task of working with classes in Python. With dataclasses, you don't need to define dunder methods like __init__
and __repr__
as they will be generated automatically by default.
For your particular use case, you can solve the issue with serializing a nested class model to a dict
/ JSON string using an optional default
callable that you can pass in to json.dumps
. For instance, you can pass a lambda (which is essentially a shorthand function) such as lambda o: o.__dict__
, which retrieves the __dict__
attribute for objects that are not inherently serializable to JSON, as shown below.
from __future__ import annotations
import json
from dataclasses import dataclass
@dataclass
class A:
my_string: str
b: list[B]
@dataclass
class B:
my_int: int
password: str
a = A('test', [B(2, '123'), B(5, '321')])
print(a)
# A(my_string='test', b=[B(my_int=2, password='123'), B(my_int=5, password='321')])
print(json.dumps(a, default=lambda o: o.__dict__))
# {"my_string": "test", "b": [{"my_int": 2, "password": "123"}, {"my_int": 5, "password": "321"}]}
But note that dataclasses also provides a helper function asdict
if you need to convert a dataclass instance to a dict
instance first.
Though, if you have a more advanced use case - for example, de-serializing JSON data to a nested class model, or mapping a dataclass field such as password
to another JSON key such as pw
as in the listed example - I'd suggest checking out a serialization library like the dataclass-wizard, which works well with a nested dataclass model as above.