init: RAG 知识库服务初始提交

- 后端 API(Flask + Gunicorn)
- RAG 引擎(混合检索 + 云端 Reranker + 引用溯源)
- 文档解析(MinerU + 多格式支持)
- Docker 生产部署配置
- 排除前端项目、敏感配置、模型文件
This commit is contained in:
lacerate551
2026-06-04 17:35:27 +08:00
commit 100d1a06eb
158 changed files with 64534 additions and 0 deletions

274
exam_pkg/api.py Normal file
View File

@@ -0,0 +1,274 @@
"""
出题与批题系统 API 蓝图
提供 REST API 接口:
- 出题:生成题目(返回 JSON 给后端)
- 批题:批阅答案(返回结果给后端)
职责边界:
- RAG 服务负责:生成题目 + 批阅答案
- 后端服务负责:审核入库 + 状态管理
使用方式:
from exam_pkg.api import exam_bp
app.register_blueprint(exam_bp, url_prefix='/exam')
"""
from flask import Blueprint, request, jsonify
import os
# 导入考试管理模块
from exam_pkg.manager import (
generate_questions_from_file,
grade_answers,
)
# 导入网关认证模块
from auth.gateway import (
require_gateway_auth, require_role, check_collection_permission, get_current_user
)
# 导入统一响应格式
from core.status_codes import EXAM_SUCCESS, GRADE_SUCCESS, EXAM_ERROR, GRADE_ERROR, BAD_REQUEST, NO_CONTENT, LLM_ERROR
from api.response_utils import success_response, error_response
# 创建蓝图
exam_bp = Blueprint('exam', __name__)
# ==================== 出题 API ====================
@exam_bp.route('/generate', methods=['POST'])
@require_gateway_auth
def api_generate_questions():
"""
🔥 新版出题接口
请求体:
{
"request_id": "uuid-optional", // 幂等性支持
"file_path": "public/产品手册.pdf",
"collection": "public_kb",
"question_types": {
"single_choice": 3,
"multiple_choice": 2,
"true_false": 2,
"fill_blank": 2,
"subjective": 1
},
"difficulty": 3,
"options": {
"include_explanation": true,
"max_source_chunks": 50
}
}
返回:
{
"success": true,
"request_id": "uuid",
"questions": [
{
"metadata": {"question_id": "...", "question_type": "...", ...},
"source_trace": {"document_name": "...", "sources": [...], ...},
"content": {"stem": "...", "data": {...}, "answer": "...", ...}
}
],
"total": 10,
"source_chunks_used": 15
}
"""
try:
data = request.json
file_path = data.get('file_path')
collection = data.get('collection')
question_types = data.get('question_types', {})
if not file_path or not collection:
return error_response("MISSING_PARAMS", BAD_REQUEST, "缺少 file_path 或 collection 参数", http_status=400)
if not question_types:
return error_response("MISSING_PARAMS", BAD_REQUEST, "缺少 question_types 参数", http_status=400)
# 获取当前用户
user = get_current_user()
if not user:
return error_response("UNAUTHORIZED", BAD_REQUEST, "未认证", http_status=401)
# 检查向量库访问权限
if not check_collection_permission(
role=user['role'],
department=user.get('department', ''),
collection_name=collection,
operation="read"
):
return error_response("FORBIDDEN", BAD_REQUEST, "权限不足", http_status=403)
# 调用新版出题接口
result = generate_questions_from_file(
file_path=file_path,
collection=collection,
question_types=question_types,
difficulty=data.get('difficulty', 3),
options=data.get('options', {}),
request_id=data.get('request_id')
)
return success_response(data=result, status_code=EXAM_SUCCESS, message="出题成功")
except Exception as e:
return error_response("EXAM_ERROR", EXAM_ERROR, str(e), http_status=500)
@exam_bp.route('/generate-smart', methods=['POST'])
@require_gateway_auth
def api_generate_smart():
"""
AI 智能出题 - 自动分析文件并决定题型和数量
与 /exam/generate 的区别:不需要传 question_typesAI 自动分析文档后决定
请求体:
{
"file_path": "public/产品手册.pdf",
"collection": "public_kb",
"difficulty": 3, // 可选,默认 3
"options": {} // 可选
}
返回:
与 /exam/generate 相同格式,额外包含 ai_analysis 字段
"""
try:
data = request.json
file_path = data.get('file_path')
collection = data.get('collection')
if not file_path or not collection:
return error_response("MISSING_PARAMS", BAD_REQUEST, "缺少 file_path 或 collection 参数", http_status=400)
# 获取当前用户
user = get_current_user()
if not user:
return error_response("UNAUTHORIZED", BAD_REQUEST, "未认证", http_status=401)
# 检查向量库访问权限
if not check_collection_permission(
role=user['role'],
department=user.get('department', ''),
collection_name=collection,
operation="read"
):
return error_response("FORBIDDEN", BAD_REQUEST, "权限不足", http_status=403)
# 1. 调用 AI 分析文件,获取推荐的题型和数量
from exam_pkg.manager import analyze_file_for_exam
ai_analysis = analyze_file_for_exam(
file_path=file_path,
collection=collection
)
question_types = ai_analysis.get('question_types', {})
if not question_types or sum(question_types.values()) == 0:
return error_response("EXAM_ERROR", EXAM_ERROR, "AI 分析后未生成有效题型配置", http_status=500)
# 2. 使用 AI 推荐的题型调用出题接口
result = generate_questions_from_file(
file_path=file_path,
collection=collection,
question_types=question_types,
difficulty=data.get('difficulty', 3),
options=data.get('options', {}),
request_id=data.get('request_id')
)
# 3. 在返回结果中添加 AI 分析信息
result['ai_analysis'] = ai_analysis
return success_response(data=result, status_code=EXAM_SUCCESS, message="AI 智能出题成功")
except Exception as e:
return error_response("EXAM_ERROR", EXAM_ERROR, str(e), http_status=500)
# ==================== 批题 API ====================
@exam_bp.route('/grade', methods=['POST'])
@require_gateway_auth
def api_grade_answers():
"""
🔥 新版批题接口
请求体:
{
"request_id": "uuid-optional",
"answers": [
{
"question_id": "uuid",
"question_type": "single_choice",
"question_content": {"answer": "B"},
"student_answer": "A",
"max_score": 2
},
{
"question_id": "uuid",
"question_type": "fill_blank",
"question_content": {"answer": [["答案1"], ["答案2"]]},
"student_answer": ["学生答案1", "学生答案2"],
"max_score": 4
},
{
"question_id": "uuid",
"question_type": "subjective",
"question_content": {
"stem": "简述...",
"data": {"scoring_points": [...]},
"answer": "参考范文..."
},
"student_answer": "学生作答内容...",
"max_score": 10
}
]
}
返回:
{
"success": true,
"request_id": "uuid",
"results": [...],
"total_score": 10.5,
"total_max_score": 16.0,
"score_rate": 65.6
}
"""
try:
data = request.json
answers = data.get('answers', [])
if not answers:
return error_response("MISSING_PARAMS", BAD_REQUEST, "缺少答案数据", http_status=400)
# 调用新版批题接口
result = grade_answers(
answers=answers,
request_id=data.get('request_id')
)
return success_response(data=result, status_code=GRADE_SUCCESS, message="批阅完成")
except Exception as e:
return error_response("GRADE_ERROR", GRADE_ERROR, str(e), http_status=500)
# ==================== 健康检查 ====================
@exam_bp.route('/health', methods=['GET'])
def api_health():
"""健康检查"""
return jsonify({
"status": "ok",
"service": "exam-api",
"version": "2.0"
})