Dataclass Dict Convert

This library converts between python dataclasses and dicts (and json).

It was created because when using the dataclasses-json library for my use case, I ran into limitations and performance issues. (There's also typed-json-dataclass but I haven't evaluated that library.)

Since creating this library, I've discovered Pydentic, which is a much more professional library, that in 99% of cases will be much more suitable than this library. So I strongly recommend using Pydentic. If you use dataclass-dict-convert, beware that it is poorly maintained, not very well documented, may contain some bugs, and has some strange edge cases where it doesn't work.

dataclass-dict-convert supports lists, optionals, dicts, enums, nested dataclasses, etc. It handles dates using RFC3339 (and enforces timezones and timezone aware datetime).

Example:

from dataclasses import dataclass
from stringcase import camelcase
from typing import Optional, List

from dataclass_dict_convert import dataclass_dict_convert


@dataclass_dict_convert(dict_letter_case=camelcase)
@dataclass(frozen=True)
class TestB:
    an_int: int
    a_str: str
    a_float: float
    a_bool: bool


@dataclass_dict_convert(dict_letter_case=camelcase)
@dataclass(frozen=True)
class Test:
    nestedClass: TestB
    nestedInOpt: Optional[TestB]
    nestedInList: List[TestB]


the_instanceB1 = TestB(1, 'foo', 0.1, True)
the_instanceB2 = TestB(2, 'bar', 0.2, False)
the_instanceB3 = TestB(3, 'baz', 0.3, True)
the_instanceB4 = TestB(4, 'huh', 0.4, False)
the_instance = Test(the_instanceB1, the_instanceB2, [the_instanceB3, the_instanceB4])
the_dict = {
    'nestedClass': {'anInt': 1, 'aStr': 'foo', 'aFloat': 0.1, 'aBool': True, },
    'nestedInOpt': {'anInt': 2, 'aStr': 'bar', 'aFloat': 0.2, 'aBool': False, },
    'nestedInList': [
        {'anInt': 3, 'aStr': 'baz', 'aFloat': 0.3, 'aBool': True, },
        {'anInt': 4, 'aStr': 'huh', 'aFloat': 0.4, 'aBool': False, },
    ],
}

expected = the_dict
actual = the_instance.to_dict()
assert actual == expected

expected = the_instance
actual = Test.from_dict(the_dict)
assert actual == expected

The library also includes:

• RFC3339 tools (the default format for converting datetime to string),
• dataclass type checking tools
• dataclass copy method generator
• dataclass multiline repr (replace auto generated repr for dataclasses by a multiline version)

Links:

• Gitlab: https://gitlab.ilabt.imec.be/wvdemeer/dataclass-dict-convert
• PyPi: https://pypi.org/project/dataclass-dict-convert/