JSONDecoder详解 - Python自定义反序列化完整指南

import json

class CustomDecoder(json.JSONDecoder):
    def __init__(self, **kwargs):
        kwargs['object_hook'] = self.object_hook
        super().__init__(**kwargs)

    def object_hook(self, obj):
        # 自定义解码逻辑
        return obj

# 使用自定义解码器
data = json.loads(json_str, cls=CustomDecoder)

import json
from datetime import datetime, date

class DateTimeDecoder(json.JSONDecoder):
    """自动还原datetime和date的解码器"""

    def __init__(self, **kwargs):
        kwargs['object_hook'] = self._hook
        super().__init__(**kwargs)

    def _hook(self, obj):
        for key, value in obj.items():
            if isinstance(value, str):
                # 尝试解析ISO日期时间
                if 'T' in value:
                    try:
                        obj[key] = datetime.fromisoformat(value)
                    except ValueError:
                        pass
                # 尝试解析纯日期
                elif len(value) == 10 and value.count('-') == 2:
                    try:
                        obj[key] = date.fromisoformat(value)
                    except ValueError:
                        pass
        return obj

# 编码
data = {"event": "会议", "time": "2024-01-15T14:30:00", "date": "2024-06-30"}
json_str = json.dumps(data, ensure_ascii=False)

# 解码
result = json.loads(json_str, cls=DateTimeDecoder)
print(f"time类型: {type(result['time']).__name__}")
print(f"time值: {result['time']}")
print(f"date类型: {type(result['date']).__name__}")
print(f"date值: {result['date']}")

time类型: datetime
time值: 2024-01-15 14:30:00
date类型: date
date值: 2024-06-30

import json
from datetime import datetime, date
from decimal import Decimal

class TypedDecoder(json.JSONDecoder):
    """基于类型标记的解码器，与UniversalEncoder配套"""

    TYPE_KEYS = {
        '__datetime__': lambda v: datetime.fromisoformat(v),
        '__date__': lambda v: date.fromisoformat(v),
        '__decimal__': lambda v: Decimal(v),
        '__set__': lambda v: set(v),
    }

    def __init__(self, **kwargs):
        kwargs['object_hook'] = self._hook
        super().__init__(**kwargs)

    def _hook(self, obj):
        # 检查是否为类型标记对象
        for type_key, converter in self.TYPE_KEYS.items():
            if type_key in obj and len(obj) == 1:
                return converter(obj[type_key])
        return obj

# 编码+解码完整循环
class TypedEncoder(json.JSONEncoder):
    def default(self, obj):
        if isinstance(obj, datetime):
            return {"__datetime__": obj.isoformat()}
        elif isinstance(obj, date):
            return {"__date__": obj.isoformat()}
        elif isinstance(obj, Decimal):
            return {"__decimal__": str(obj)}
        elif isinstance(obj, set):
            return {"__set__": list(obj)}
        return super().default(obj)

original = {
    "created": datetime(2024, 1, 15, 14, 30),
    "deadline": date(2024, 6, 30),
    "price": Decimal("99.99"),
    "tags": {"python", "json"},
}

# 序列化
json_str = json.dumps(original, cls=TypedEncoder, ensure_ascii=False)
print("JSON:", json_str[:80], "...")

# 反序列化
restored = json.loads(json_str, cls=TypedDecoder)
print(f"\ncreated类型: {type(restored['created']).__name__}")
print(f"price类型: {type(restored['price']).__name__}")
print(f"tags类型: {type(restored['tags']).__name__}")

JSON: {"created": {"__datetime__": "2024-01-15T14:30:00"}, "deadline": {"__date__": "2024-0 ...

created类型: datetime
price类型: Decimal
tags类型: set

import json

def parse_json_stream(stream_str):
    """解析包含多个JSON对象的字符串流"""
    decoder = json.JSONDecoder()
    results = []
    idx = 0

    while idx < len(stream_str):
        # 跳过空白
        while idx < len(stream_str) and stream_str[idx].isspace():
            idx += 1

        if idx >= len(stream_str):
            break

        try:
            obj, end = decoder.raw_decode(stream_str, idx)
            results.append(obj)
            idx = end
        except json.JSONDecodeError as e:
            print(f"解析错误位置 {idx}: {e}")
            break

    return results

# 测试：连续的JSON对象
stream = '{"a": 1}  {"b": 2}  {"c": 3}'
results = parse_json_stream(stream)

print(f"解析出 {len(results)} 个对象:")
for i, obj in enumerate(results, 1):
    print(f"  对象{i}: {obj}")

解析出 3 个对象:
  对象1: {'a': 1}
  对象2: {'b': 2}
  对象3: {'c': 3}