Serializer

Pluggable value encoding/decoding for all cache layers. Default is PickleSerializer. Switch to JsonSerializer for Redis backends where untrusted data may be at rest.

Built-in Serializers

Class	Format	Supports	Use when
PickleSerializer	Python pickle	Any Python object	Default — most flexible
JsonSerializer	JSON (UTF-8 bytes)	str, int, float, bool, list, dict, None	Redis + security-conscious deployments

---

Usage

Default (pickle)

from chengeta_ai import ResponseCache, CacheManager, InMemoryBackend, CacheKeyBuilder

manager = CacheManager(backend=InMemoryBackend(), key_builder=CacheKeyBuilder())
cache = ResponseCache(manager)  # uses PickleSerializer by default

JSON serializer (safer for Redis)

from chengeta_ai import JsonSerializer, ResponseCache

cache = ResponseCache(manager, serializer=JsonSerializer())

Custom serializer

Implement the Serializer protocol — two methods required:

from chengeta_ai import Serializer
import msgpack

class MsgpackSerializer:
    def dumps(self, value) -> bytes:
        return msgpack.packb(value)

    def loads(self, data: bytes):
        return msgpack.unpackb(data)

# Use with any layer
cache = ResponseCache(manager, serializer=MsgpackSerializer())

---

Serializer Protocol

from chengeta_ai import Serializer
from typing import Any

class Serializer(Protocol):
    def dumps(self, value: Any) -> bytes: ...
    def loads(self, data: bytes) -> Any: ...

Serializer is @runtime_checkable — verify with isinstance(my_ser, Serializer).

---

Security Note

PickleSerializer deserializes arbitrary Python objects. For Redis backends shared across services or exposed to untrusted input, use JsonSerializer to limit the attack surface. JSON is also more interoperable if you need to inspect cache values from other languages.