feat: 新增多格式文件上传入库与认证体系

- 新增 JWT 认证模块，支持登录/注册/用户管理
- 新增文件上传接口，支持 .txt/.md/.html/.pdf/.docx 等格式解析入库
- 新增检索结果 AI 总结功能
- 新增文本去重缓存机制
- 新增全局认证夹具简化测试
- 新增配置项与环境变量支持
- 完善文档与测试覆盖

app/models/auth.py

@@ -0,0 +1,42 @@
+"""认证相关数据模型"""
+
+from datetime import datetime
+
+from pydantic import BaseModel, Field
+
+
+class LoginRequest(BaseModel):
+    """登录请求"""
+
+    username: str = Field(description="用户名")
+    password: str = Field(description="明文密码")
+
+
+class RegisterRequest(BaseModel):
+    """注册请求"""
+
+    username: str = Field(min_length=3, max_length=32, description="用户名（3-32 字符）")
+    password: str = Field(min_length=6, max_length=128, description="密码（6-128 字符）")
+
+
+class AuthUser(BaseModel):
+    """对外暴露的用户信息（不含密码）"""
+
+    username: str = Field(description="用户名")
+    role: str = Field(default="user", description="角色：admin | user")
+    created_at: datetime = Field(description="创建时间")
+
+
+class StoredUser(AuthUser):
+    """存储层用户：含密码哈希，仅内部使用，不对外暴露"""
+
+    hashed_password: str = Field(description="bcrypt 密码哈希")
+
+
+class TokenResponse(BaseModel):
+    """登录成功返回的 token 信息"""
+
+    access_token: str = Field(description="JWT access token")
+    token_type: str = Field(default="bearer", description="token 类型")
+    expires_in: int = Field(description="token 有效期（秒）")
+    user: AuthUser = Field(description="登录用户信息")

app/models/document.py

@@ -49,3 +49,4 @@ class IngestionResult(BaseModel):
     chunks_count: int = Field(default=0, description="写入的 chunk 数量")
     tags: list[str] = Field(default_factory=list, description="附加分类标签")
     category_confidence: float = Field(default=0.0, description="主类目分类置信度")
+    deduplicated: bool = Field(default=False, description="是否命中去重复用旧文档（True 时 document_id 为既有文档 ID）")

app/models/search.py

@@ -8,6 +8,7 @@ class SearchRequest(BaseModel):
 
     query: str = Field(description="查询文本")
     top_k: int | None = Field(default=None, description="返回结果数，为空时使用 settings.retrieval_final_k")
+    summarize: bool = Field(default=False, description="是否对检索结果生成 AI 总结")
 
 
 class SearchHit(BaseModel):
@@ -21,6 +22,17 @@ class SearchHit(BaseModel):
     doc_summary: str = Field(default="", description="L1 文档总结，仅用于上下文标注")
 
 
+class ExtractedInfo(BaseModel):
+    """AI 从 query 中提取的关键信息"""
+
+    rewrite: str = Field(default="", description="改写后的 query")
+    keywords: list[str] = Field(default_factory=list, description="关键词")
+    entities: list[str] = Field(default_factory=list, description="实体（人名/产品/技术等）")
+    intent: str = Field(default="", description="查询意图分类")
+    time_range: str = Field(default="", description="时间范围，空表示无")
+    categories: list[str] = Field(default_factory=list, description="命中类目名（含未达阈值的候选）")
+
+
 class SearchResponse(BaseModel):
     """检索响应"""
 
@@ -28,3 +40,5 @@ class SearchResponse(BaseModel):
     hits: list[SearchHit] = Field(default_factory=list, description="命中结果列表")
     routed_categories: list[str] = Field(default_factory=list, description="query 路由命中的类目")
     fallback: bool = Field(default=False, description="是否走了全库兜底路径")
+    extracted_info: ExtractedInfo | None = Field(default=None, description="AI 提取的 query 关键信息")
+    summary: str | None = Field(default=None, description="检索结果 AI 总结，仅 summarize=true 时返回")