查看LangChain源码时,发现Literal:
python
class Document(BaseMedia):
"""Class for storing a piece of text and associated metadata.
!!! note
`Document` is for **retrieval workflows**, not chat I/O. For sending text
to an LLM in a conversation, use message types from `langchain.messages`.
Example:
```python
from langchain_core.documents import Document
document = Document(
page_content="Hello, world!", metadata={"source": "https://example.com"}
)
```
"""
page_content: str
"""String text."""
type: Literal["Document"] = "Document"
作用:静态代码检查时,如果type为其他值,即!="Document",会发生报错。 注:在不实际运行程序(不执行代码)的情况下,通过分析源代码的文本结构来找出潜在的错误。
进一步,LangChain 框架中的深层作用
- 序列化时的"防伪标签" 当 LangChain 将 Document 对象转换成 JSON 字符串时,type: Literal"Document" 保证了输出的 JSON 里一定会包含这个标签 如下:
python
print(f"转换格式后:\n{json.dumps(document1_json, ensure_ascii=False, indent=2)}")
# 转换格式后,{
# "lc": 1,
# "type": "constructor",
# "id": [
# "langchain",
# "schema",
# "document",
# "Document"
# ],
# "kwargs": {
# "metadata": {
# "author": "张三",
# "page": 10
# },
# "page_content": "LangChain 是一个用于开发大语言模型应用的框架。",
# "type": "Document"
# }
# }
- 配合 Pydantic 的运行时校验 Pydantic 会直接抛出验证错误,防止脏数据污染系统
例子:
python
from pydantic import BaseModel
from typing import Literal
class MyClass(BaseModel):
type: Literal["red"] = "red"
obj = MyClass()
obj.type = "Image" # ❌ Pydantic 会在运行时抛出 ValidationError