Install

pip install dataclasses_ujson

JSON with DataClasses

The library provides a simple API for decoding JSON to dataclasses. You can use nested dataclasses. The library uses ujson for better performance. There is support for validate simply type: str, int, float, bool, Optional Minimal python version: python 3.7

Examples

from typing import List, Dict
from dataclasses import dataclass
from dataclasses_ujson.dataclasses_ujson import UJsonMixin

json_string = """
{"a": 1, "b": [{"x": 1}, {"x": 2}], "c": {"x": 1}}
"""

@dataclass(frozen=True)
class JsonDict(UJsonMixin):
    x: Dict[str, int]

@dataclass(frozen=True)
class JsonClass(UJsonMixin):
    a: int
    b: List[JsonDict]
    c: Dict[str, int]

data = JsonClass.loads(json_string)

print(data.c["x"])
print(list(data.b)[0].x)

new_dict = data.to_serializable()
print(new_dict)

All lists will be returned as generators

from typing import List, Dict
from dataclasses import dataclass
from dataclasses_ujson.dataclasses_ujson import UJsonMixin

json_string = """
[{"x": 1}, {"x": 2}]
"""

@dataclass(frozen=True)
class JsonDict(UJsonMixin):
    x: Dict[str, int]

data = JsonDict.loads(json_string, many=True)

print(data) # generator object UJsonMixin
print(list(data)) # list of JsonDict

Performance:

Libraries were compared:

• The default library json of python
• The library ujson https://github.com/esnme/ultrajson
• The library dataclasses-json https://github.com/lidatong/dataclasses-json

The script is placed in repository (bench_marks.py):

Name of library	Results
json	0.62s
ujson	0.53s
dataclasses-ujson	1.24s (python3.10)
dataclasses-json	21.6s

if generator will be returned (using flag many=true)

Name of library	Results
json	0.62s
ujson	0.54s
dataclasses-ujson	0.59s
dataclasses-json	-