nested dataclass serialization

• lets you (de)serialize nested dataclasses from and to json
• uses custom json-encoder for recursive serialization
• uses importlib for deserialization (instantiation of the dataclass-objects)

minimal example

@dataclass
class Bar:
    x: str


@dataclass
class Foo:
    bars: list[Bar]

bar = Bar(x="test")
foo = Foo(bars=[bar, bar])

json_str = serialize_dataclass(foo)

# this json_str looks like this:
{
    "_target_": "test_dataclass_serialization.Foo",
    "_id_": "a22ddd36-<some-long-id-here>d2290929",
    "bars": [
        {
            "_target_": "test_dataclass_serialization.Bar",
            "_id_": "a22ddd36-<some-long-id-here>ebb0ab925b1bd977208e4",
            "x": "test",
        },
        {
            "_target_": "test_dataclass_serialization.Bar",
            "_id_": "a22ddd36-<some-long-id-here>ebb0ab925b1bd977208e4",
            "x": "test",
        },
    ],
}
# somewhere in a different python-process/container/server
des_foo = deserialize_dataclass(json_str) # getting back the dataclass-object
assert id(des_foo.bars[0]) == id(des_foo.bars[1]) # just to show that deserialization works

similar libs

• madman-bob: looks like you need to explicitly provide the python-class for deserialization: JSONSerializer.deserialize(InventoryItem, {'name': 'Apple', 'unit_ ...