"""Schemas for the keyword-based subjective grading mini-model.
中文：主观题评分的请求/响应模型，校验题目与作答字段并规范输出结构。
"""
from __future__ import annotations
from typing import List
from pydantic import BaseModel, Field
class GradeAnswerRequest(BaseModel):
question: str = Field(..., description="题目内容")
correct_answer: str = Field(..., description="标准答案")
student_answer: str = Field(..., description="学生作答")
class GradeAnswerResponse(BaseModel):
is_correct: bool = Field(..., description="是否判定为正确")
similarity: float = Field(..., ge=0.0, le=1.0, description="0～1 之间的覆盖率")
matched_keywords: List[str] = Field(default_factory=list, description="命中的关键词")
missing_keywords: List[str] = Field(default_factory=list, description="缺失的关键词")
feedback: str = Field(..., description="自然语言反馈")
