typedjson

JSON decoding for Python with type hinting (PEP 484).

Requirements and Restrictions

• Python >= 3.6
• Mypy <= 0.770
• Use non-generic or parameterized class to decode JSON.
• Use type hints without forward references.

Features

• Support decoding types as below:
  • primitive types like str, int, float, bool and None.
  • Union and Optional.
  • homogeneous and heterogeneous Tuple and List.
  • variable-length Tuple.
  • non-generic and parameterized dataclasses.
• Support API like json.load and json.loads.

Example

from typing import Optional

import typedjson
from dataclasses import dataclass


@dataclass(frozen=True)
class NameJson:
    first: str
    last: Optional[str]


@dataclass(frozen=True)
class CatJson:
    id: str
    age: int
    name: Optional[NameJson]


json = {
    'id': 'test-cat',
    'age': 13,
    'name': {
        'first': 'Jiji',
    },
}

print(typedjson.decode(CatJson, json))  # Output: CatJson(id='test-cat', age=13, name=NameJson(first='Jiji', last=None))

print(typedjson.decode(CatJson, {}))  # Output: DecodingError(TypeMismatch(('id',)))

Please refer to test codes for more detail.

Contributions

Please read CONTRIBUTING.md.

TODO

• Prohibit decoding Set and Dict explicitly.
• Provide the API document.
• Explain why typedjson uses undocumented APIs.
• Explain what typedjson resolves.
• Improve API to dump like json.dump and json.dumps.
  • Provide mypy plugin to check whether the class is encodable as JSON or not with @typedjson.encodable decorator.
• Improve the peformance of typedjson.decode.
• Support type hints with forward reference.
• Support TypedDict.