// 传统成绩发布流程（伪代码）

function traditionalGradePublish(excelFile):
    grades = readExcel(excelFile)
    for each student in grades:
        // 老师手动复制成绩 → 打开聊天窗口 → 逐条发送
        sendToWeChat(student.phone, student.name + "成绩：" + student.score)
        // 手动计算平均分、排名
        calculateStats(grades)
        // 手动制作Excel图表
        exportToChart(grades)
    return "耗时2小时，期间还要应对家长追问"

 AI查分助手核心模块 - 极简实现
import pandas as pd
from sklearn.feature_extraction.text import TfidfVectorizer
from sklearn.metrics.pairwise import cosine_similarity

class SimpleAIGradingAssistant:
    """AI查分助手核心类 - 演示自动评分与成绩分析"""
    
    def __init__(self, standard_answers: dict):
         标准答案库
        self.standard_answers = standard_answers
         TF-IDF向量化器，用于计算答案相似度
        self.vectorizer = TfidfVectorizer()
         预计算标准答案的向量
        self.standard_vectors = {}
        for qid, answer in standard_answers.items():
            self.standard_vectors[qid] = self.vectorizer.fit_transform([answer])
    
    def score_subjective_answer(self, question_id: str, student_answer: str) -> dict:
        """评分主观题（基于相似度算法，AI判分引擎的核心逻辑）"""
        if question_id not in self.standard_answers:
            return {"score": 0, "reason": "题目不存在"}
        
         步骤1：向量化学生答案
        student_vec = self.vectorizer.transform([student_answer])
         步骤2：计算与标准答案的余弦相似度
        similarity = cosine_similarity(student_vec, self.standard_vectors[question_id])[0][0]
         步骤3：根据相似度映射得分（0-10分制）
        raw_score = similarity  10
        score = min(10, round(raw_score, 1))
        
         步骤4：生成判分理由（AI的可解释性）
        if score >= 9:
            reason = "答案与标准答案高度匹配，表述精准"
        elif score >= 7:
            reason = "答案核心要点正确，表达较为完整"
        elif score >= 5:
            reason = "答案方向正确，但关键点表达不够充分"
        else:
            reason = "答案与标准答案偏离较大，建议加强学习"
        
        return {"score": score, "similarity": round(similarity, 3), "reason": reason}
    
    def generate_report(self, student_scores: dict) -> dict:
        """生成成绩分析报告（AI查分助手的统计与分析功能）"""
        scores = list(student_scores.values())
        avg_score = sum(scores) / len(scores)
        max_score = max(scores)
        min_score = min(scores)
        
         自动生成AI智能评语
        if avg_score >= 8:
            comment = "优秀！各项表现均超出预期，继续保持。"
        elif avg_score >= 6:
            comment = "良好，整体掌握较为扎实，建议在薄弱环节多下功夫。"
        else:
            comment = "有待提高，建议系统复习知识点，查漏补缺。"
        
        return {
            "average": round(avg_score, 2),
            "highest": max_score,
            "lowest": min_score,
            "ai_comment": comment,
            "total_students": len(scores)
        }
    
    def batch_process(self, student_answers: dict) -> pd.DataFrame:
        """批量处理所有学生的答卷，生成成绩表"""
        results = []
        for student_id, answers in student_answers.items():
            student_result = {"student_id": student_id}
            total_score = 0
            for qid, answer in answers.items():
                score_result = self.score_subjective_answer(qid, answer)
                student_result[f"Q{qid}_score"] = score_result["score"]
                student_result[f"Q{qid}_reason"] = score_result["reason"]
                total_score += score_result["score"]
            student_result["total_score"] = total_score
            student_result["ai_comment"] = self.generate_report({qid: total_score})["ai_comment"]
            results.append(student_result)
        return pd.DataFrame(results)

 ========== 使用示例 ==========
 配置标准答案
standard = {
    "1": "人工智能的核心是让机器模拟人类的认知功能，包括学习、推理和感知。",
    "2": "深度学习的本质是通过多层神经网络自动提取数据的层次化特征表示。"
}

 创建AI查分助手实例
assistant = SimpleAIGradingAssistant(standard)

 模拟学生答卷
student_answers = {
    "S001": {"1": "人工智能是让机器模拟人类智能的技术，包括学习能力和推理能力。", 
             "2": "深度学习用神经网络从数据中提取特征，层数越多提取越抽象。"},
    "S002": {"1": "AI就是机器人。", 
             "2": "深度学习是机器学习的一个分支。"}
}

 批量批改并生成成绩表
results_df = assistant.batch_process(student_answers)
print("=== AI查分助手批改结果 ===")
print(results_df[["student_id", "total_score", "ai_comment"]])