如何修改Pydantic的GrammarSentence模型以适配空响应场景？

方案1：修改字段类型，允许列表包含None元素 #

直接调整corrected_sentences的类型定义，让它接受dict[str, str]或None的列表，这样API返回[None]时就不会触发验证错误。

from pydantic import BaseModel
from typing import List, Dict

class GrammarSentence(BaseModel):
    """
    Model for the response containing a list of sentences with grammar mistake.
    """
    corrected_sentences: List[Dict[str, str] | None]

方案2：用验证器自动处理None元素 #

如果希望返回的结构更统一（避免列表里出现None），可以添加Pydantic验证器，在数据验证前把列表中的None替换成空字典，或者直接过滤掉None。

替换None为空字典 #

from pydantic import BaseModel, validator
from typing import List, Dict

class GrammarSentence(BaseModel):
    """
    Model for the response containing a list of sentences with grammar mistake.
    """
    corrected_sentences: List[Dict[str, str]]

    @validator("corrected_sentences", pre=True)
    def replace_none_with_empty_dict(cls, value):
        return [item if item is not None else {} for item in value]

过滤None元素返回空列表 #

如果无语法错误时希望返回空列表而非带空元素的列表，修改验证器逻辑即可：

from pydantic import BaseModel, validator
from typing import List, Dict

class GrammarSentence(BaseModel):
    """
    Model for the response containing a list of sentences with grammar mistake.
    """
    corrected_sentences: List[Dict[str, str]]

    @validator("corrected_sentences", pre=True)
    def filter_out_none(cls, value):
        return [item for item in value if item is not None]

方案3：将[None]自动转换为空列表 #

如果业务上无错误时应该返回空列表，可以让验证器直接把[None]替换成[]，同时给字段设置默认空列表：

from pydantic import BaseModel, validator
from typing import List, Dict

class GrammarSentence(BaseModel):
    """
    Model for the response containing a list of sentences with grammar mistake.
    """
    corrected_sentences: List[Dict[str, str]] = []

    @validator("corrected_sentences", pre=True)
    def handle_none_list(cls, value):
        return [] if value == [None] else value

内容的提问来源于stack exchange，提问作者Jay Nair