from langchain.schema import BaseMemory
class ChatMemory(BaseMemory):
    def __init__(self, user_id: UUID, type: str):
        self.user_id = user_id
        self.type = type
class AnotherMem(ChatMemory):
    def __init__(self, user_id: UUID, type: str):
        super().__init__(user_id, type)

<details>
<summary>英文:</summary>
    from langchain.schema import BaseMemory
    class ChatMemory(BaseMemory):
       def __init__(self, user_id: UUID, type: str):
        self.user_id = user_id
        self.type = type
       # implemented abstract methods
    class AnotherMem(ChatMemory):
        def __init__(self, user_id: UUID, type: str):
            super().__init__(user_id, type)
This seems simple enough - but I get an error: `ValueError: &quot;AnotherMem&quot; object has no field &quot;user_id&quot;`. What am I doing wrong?
Note that [BaseMemory][1] is an interface. 
  [1]: https://github.com/hwchase17/langchain/blob/db45970a66f39a32f2cdd83e7bde26a404efad7b/langchain/schema.py#L188
</details>
# 答案1
**得分**: 0
看起来`BaseMemory`来自[langchain][1]被定义为[pydantic][2]模型，它对定义实例属性有严格的规定。不要使用构造函数方法，而是使用标准的 pydantic 语法，在子类上声明预期的实例属性。

...

  [1]: https://pypi.org/project/langchain/
  [2]: https://pypi.org/project/pydantic/
<details>
<summary>英文:</summary>
It looks like `BaseMemory` from [langchain][1] is defined as a [pydantic][2] model, which has strict rules for defining instance attributes. Instead of using a constructor method, use the standard pydantic syntax of declaring expected instance attributes on the child class.

...

  [1]: https://pypi.org/project/langchain/
  [2]: https://pypi.org/project/pydantic/
</details>