Skip to main content
Version: 1.0.0

Type Registration

This page covers Python type registration APIs. Use Configuration for strict-mode policy, max-depth limits, and trusted-data guidance.

Type Registration​

Register xlang classes by type name so other runtimes can resolve the same schema identity:

from dataclasses import dataclass
import pyfory

fory = pyfory.Fory(xlang=True, strict=True)

@dataclass
class User:
    name: str
    age: pyfory.Int32

fory.register(User, typename="example.User")

For Python native mode, numeric type IDs are the compact same-language registration path:

import pyfory

fory = pyfory.Fory(xlang=False, strict=True)
fory.register(MyClass, type_id=100)

Registration Patterns​

Use the registration form that matches the payload contract:

# Xlang: stable namespace/type-name identity
fory.register(MyClass, typename="com.example.MyClass")

# Native mode: compact numeric identity
fory.register(MyClass, type_id=100)

# Custom serializer
fory.register(MyClass, type_id=100, serializer=MySerializer(fory.type_resolver, MyClass))

# Batch registration
type_id = 100
for model_class in [User, Order, Product, Invoice]:
    fory.register(model_class, type_id=type_id)
    type_id += 1

Strict Mode Relationship​

With strict=True, deserialization accepts only registered types. Register all application classes before serializing or deserializing payloads, and keep the same registration IDs or names on every peer that shares those payloads.