philiprehberger-safe-json

JSON encoder that handles datetime, Decimal, UUID, dataclasses, and sets without crashing.

Installation

pip install philiprehberger-safe-json

Usage

from philiprehberger_safe_json import dumps, loads

data = {
    "created": datetime(2026, 3, 13, 12, 0, 0),
    "price": Decimal("19.99"),
    "id": UUID("12345678-1234-5678-1234-567812345678"),
    "tags": {"beta", "release"},
}

json_string = dumps(data)
parsed = loads(json_string)

Datetime and Date

from datetime import datetime, date
from philiprehberger_safe_json import dumps

dumps({"timestamp": datetime(2026, 1, 15, 9, 30, 0)})
# '{"timestamp": "2026-01-15T09:30:00"}'

dumps({"day": date(2026, 1, 15)})
# '{"day": "2026-01-15"}'

Decimal

from decimal import Decimal
from philiprehberger_safe_json import dumps

dumps({"price": Decimal("9.99")})
# '{"price": 9.99}'

dumps({"price": Decimal("9.99")}, decimal_as_string=True)
# '{"price": "9.99"}'

UUID

from uuid import UUID
from philiprehberger_safe_json import dumps

dumps({"id": UUID("abcdef01-2345-6789-abcd-ef0123456789")})
# '{"id": "abcdef01-2345-6789-abcd-ef0123456789"}'

Dataclasses

from dataclasses import dataclass
from philiprehberger_safe_json import dumps

@dataclass
class User:
    name: str
    age: int

dumps({"user": User(name="Alice", age=30)})
# '{"user": {"name": "Alice", "age": 30}}'

Sets and Frozensets

from philiprehberger_safe_json import dumps

dumps({"tags": {"c", "a", "b"}})
# '{"tags": ["a", "b", "c"]}'

Custom Type Encoders

from philiprehberger_safe_json import dumps, register_encoder

class Money:
    def __init__(self, amount: int, currency: str) -> None:
        self.amount = amount
        self.currency = currency

register_encoder(Money, lambda m: {"amount": m.amount, "currency": m.currency})

dumps({"payment": Money(1000, "USD")})
# '{"payment": {"amount": 1000, "currency": "USD"}}'

Circular Reference Detection

from philiprehberger_safe_json import dumps, CircularReferenceError

data = {"key": "value"}
data["self"] = data  # circular reference

dumps(data, detect_cycles=True)
# Raises CircularReferenceError

Safe Loads with Auto-Parsing

from philiprehberger_safe_json import safe_loads

result = safe_loads('{"created": "2026-03-13T14:30:00", "price": 19.99}')
# result["created"] -> datetime(2026, 3, 13, 14, 30, 0)
# result["price"] -> Decimal("19.99")

result = safe_loads('{"day": "2026-03-13"}')
# result["day"] -> date(2026, 3, 13)

Using SafeJsonEncoder Directly

import json
from philiprehberger_safe_json import SafeJsonEncoder

json.dumps({"key": some_value}, cls=SafeJsonEncoder)

API

Function / Class	Description
SafeJsonEncoder	json.JSONEncoder subclass that handles datetime, date, Decimal, UUID, dataclass, set, frozenset, bytes, Enum, and Path
SafeJsonEncoder.decimal_as_string	Class attribute; when True, Decimal values serialize as strings instead of floats (default: False)
dumps(obj, *, decimal_as_string=False, detect_cycles=False, **kwargs)	Serialize to JSON string using SafeJsonEncoder. Set detect_cycles=True to raise CircularReferenceError on circular refs
loads(s, **kwargs)	Deserialize a JSON string. Pass-through to json.loads for API symmetry
safe_loads(s, *, parse_dates=True, parse_decimals=True, **kwargs)	Deserialize with auto-parsing of ISO date strings to datetime/date and numeric values to Decimal
register_encoder(type_class, handler_fn)	Register a custom encoder for a specific type without subclassing
clear_encoders()	Remove all registered custom encoders
CircularReferenceError	Exception raised when a circular reference is detected during serialization

Development

pip install -e .
python -m pytest tests/ -v

Support

If you find this project useful:

⭐ Star the repo

🐛 Report issues

💡 Suggest features

❤️ Sponsor development

🌐 All Open Source Projects

💻 GitHub Profile

🔗 LinkedIn Profile

License

MIT