feat(admin): Add user management and upgrade to module permission system

Features - User Management (Phase 4.1):
- Database: Add user_modules table for fine-grained module permissions
- Database: Add 4 user permissions (view/create/edit/delete) to role_permissions
- Backend: UserService (780 lines) - CRUD with tenant isolation
- Backend: UserController + UserRoutes (648 lines) - 13 API endpoints
- Backend: Batch import users from Excel
- Frontend: UserListPage (412 lines) - list/filter/search/pagination
- Frontend: UserFormPage (341 lines) - create/edit with module config
- Frontend: UserDetailPage (393 lines) - details/tenant/module management
- Frontend: 3 modal components (592 lines) - import/assign/configure
- API: GET/POST/PUT/DELETE /api/admin/users/* endpoints

Architecture Upgrade - Module Permission System:
- Backend: Add getUserModules() method in auth.service
- Backend: Login API returns modules array in user object
- Frontend: AuthContext adds hasModule() method
- Frontend: Navigation filters modules based on user.modules
- Frontend: RouteGuard checks requiredModule instead of requiredVersion
- Frontend: Remove deprecated version-based permission system
- UX: Only show accessible modules in navigation (clean UI)
- UX: Smart redirect after login (avoid 403 for regular users)

Fixes:
- Fix UTF-8 encoding corruption in ~100 docs files
- Fix pageSize type conversion in userService (String to Number)
- Fix authUser undefined error in TopNavigation
- Fix login redirect logic with role-based access check
- Update Git commit guidelines v1.2 with UTF-8 safety rules

Database Changes:
- CREATE TABLE user_modules (user_id, tenant_id, module_code, is_enabled)
- ADD UNIQUE CONSTRAINT (user_id, tenant_id, module_code)
- INSERT 4 permissions + role assignments
- UPDATE PUBLIC tenant with 8 module subscriptions

Technical:
- Backend: 5 new files (~2400 lines)
- Frontend: 10 new files (~2500 lines)
- Docs: 1 development record + 2 status updates + 1 guideline update
- Total: ~4900 lines of code

Status: User management 100% complete, module permission system operational

docs/03-业务模块/ASL-AI智能文献/02-技术设计/01-数据库设计.md

@@ -1,14 +1,17 @@
-# AI<EFBFBD>箄<EFBFBD><EFBFBD><EFBFBD>讃璅∪<EFBFBD> - <20>唳旿摨栞挽霈?
-> **<2A><>﹝<EFBFBD><EFB99D>𧋦嚗?* v3.0  
-> **<EFBFBD>𥕦遣<EFBFBD>交<EFBFBD>嚗?* 2025-10-29  
-> **蝏湔擪<EFBFBD><EFBFBD><EFBFBD>** AI<41>箄<EFBFBD><E7AE84><EFBFBD>讃撘<E8AE83><E69298>穃𣪧<E7A983>? 
-> **<EFBFBD><EFBFBD><EFBFBD>擧凒<EFBFBD>堆<EFBFBD>** 2025-11-22嚗㇄ay 4嚗𡁜<E59A97><F0A1819C><EFBFBD><EFBFBD>蝑𥟇㺭<F0A59F87>桀<EFBFBD>霈曇恣嚗? 
-> **<EFBFBD>湔鰵霂湔<EFBFBD>嚗?* <20>啣<EFBFBD><E595A3>冽<EFBFBD>憭滨<E686AD><E6BBA8>詨<EFBFBD>銵剁<E98AB5>`AslLiterature`<60>拙<EFBFBD><E68B99><EFBFBD>AslFulltextScreeningTask`<60><>AslFulltextScreeningResult`嚗?
+# AI智能文献模块 - 数据库设计
+
+> **文档版本：** v3.0  
+> **创建日期：** 2025-10-29  
+> **维护者：** AI智能文献开发团队  
+> **最后更新：** 2025-11-22（Day 4：全文复筛数据库设计）  
+> **更新说明：** 新增全文复筛相关表（`AslLiterature`扩展、`AslFulltextScreeningTask`、`AslFulltextScreeningResult`）
+
 ---
 
 ## 📋 文档说明
 
-<EFBFBD>祆<EFBFBD>獢<EFBFBD><EFBFBD>餈蚊I<EFBFBD>箄<EFBFBD><EFBFBD><EFBFBD>讃璅∪<EFBFBD><EFBFBD><EFBFBD>㺭<EFBFBD>桀<EFBFBD>霈曇恣嚗<EFBFBD><EFBFBD><EFBFBD>祆㺭<EFBFBD>株”蝏𤘪<EFBFBD><EFBFBD><EFBFBD><EFBFBD>蝟餉挽霈～<EFBFBD><EFBFBD>揣撘閗挽霈∠<EFBFBD><EFBFBD>?
+本文档描述AI智能文献模块的数据库设计，包括数据表结构、关系设计、索引设计等。
+
 **技术栈**:
 - 数据库：PostgreSQL 16+
 - ORM：Prisma
@@ -17,30 +20,36 @@
 
 ---
 
-## <EFBFBD><EFBFBD>儭?Schema<EFBFBD>嗆<EFBFBD>
+## 🏗️ Schema架构
+
+ASL模块使用独立的 `asl_schema` 进行数据隔离，确保模块独立性和数据安全。
 
-ASL璅∪<EFBFBD>雿輻鍂<EFBFBD>祉<EFBFBD><EFBFBD>?`asl_schema` 餈𥡝<E9A488><F0A5A19D>唳旿<E594B3>𠉛氖嚗𣬚＆靽脲芋<E884B2>㛖𡠺蝡𧢲<E89DA1>批<EFBFBD><E689B9>唳旿摰匧<E691B0><E58CA7>?
 ```
 platform_schema
-  <EFBFBD>婙<EFBFBD><EFBFBD><EFBFBD> users (<EFBFBD>冽<EFBFBD>銵?
-       <EFBFBD>?asl_schema
-  <20>鎿<EFBFBD><E98EBF><EFBFBD> screening_projects (蝑偦<E89D91>厰★<E58EB0>?
+  └── users (用户表)
+       ↓
+asl_schema
+  ├── screening_projects (筛选项目)
   ├── literatures (文献条目)
   ├── screening_results (标题初筛结果)
   ├── screening_tasks (标题初筛任务)
-  <EFBFBD>鎿<EFBFBD><EFBFBD><EFBFBD> fulltext_screening_tasks (<EFBFBD>冽<EFBFBD>憭滨<EFBFBD>隞餃𦛚) 潃?Day 4<EFBFBD>啣<EFBFBD>
-  <EFBFBD>婙<EFBFBD><EFBFBD><EFBFBD> fulltext_screening_results (<EFBFBD>冽<EFBFBD>憭滨<EFBFBD>蝏𤘪<EFBFBD>) 潃?Day 4<EFBFBD>啣<EFBFBD>
+  ├── fulltext_screening_tasks (全文复筛任务) ⭐ Day 4新增
+  └── fulltext_screening_results (全文复筛结果) ⭐ Day 4新增
 ```
 
-**v3.0 <EFBFBD>湔鰵霂湔<EFBFBD>嚗?025-11-22嚗?*嚗?- <20>?<3F>拙<EFBFBD> `literatures` 銵剁<E98AB5><E58981>舀<EFBFBD><E88880>冽<EFBFBD><E586BD>笔𦶢<E7AC94>冽<EFBFBD>蝞∠<E89D9E><E288A0><EFBFBD>DF摮睃<E691AE><E79D83><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD>摰孵<E691B0><E5ADB5>?- <20>?<3F>啣<EFBFBD> `fulltext_screening_tasks` 銵剁<E98AB5>蝞∠<E89D9E><E288A0>冽<EFBFBD>憭滨<E686AD><E6BBA8>孵<EFBFBD><E5ADB5><EFBFBD>遙<EFBFBD>?- <20>?<3F>啣<EFBFBD> `fulltext_screening_results` 銵剁<E98AB5>摮睃<E691AE>12摮埈挾霂<E68CBE>摯蝏𤘪<E89D8F>
-- <EFBFBD>?蝚血<E89D9A>鈭穃<E988AD><E7A983>蠘<EFBFBD><E8A098><EFBFBD><EFBFBD><EFBFBD>冽<EFBFBD><E586BD><EFBFBD>捆摮睃<E691AE>撘閧鍂<E996A7>屸<EFBFBD><E5B1B8>湔𦻖摮睃<E691AE>
+**v3.0 更新说明（2025-11-22）**：
+- ✅ 扩展 `literatures` 表：支持全文生命周期管理、PDF存储、全文内容引用
+- ✅ 新增 `fulltext_screening_tasks` 表：管理全文复筛批处理任务
+- ✅ 新增 `fulltext_screening_results` 表：存储12字段评估结果
+- ✅ 符合云原生规范：全文内容存储引用而非直接存储
 
 ---
 
-## <EFBFBD><EFBFBD>儭?<3F>詨<EFBFBD><E8A9A8>唳旿銵?
+## 🗄️ 核心数据表
+
 ### 1. 筛选项目表 (screening_projects)
 
-**Prisma璅∪<EFBFBD><EFBFBD>?*: `AslScreeningProject`  
+**Prisma模型名**: `AslScreeningProject`  
 **表名**: `asl_schema.screening_projects`
 
 ```prisma
@@ -55,19 +64,23 @@ model AslScreeningProject {
   picoCriteria      Json     @map("pico_criteria")
   // ⚠️ 格式兼容性说明：
   // 前端使用: { P, I, C, O, S }
-  // <EFBFBD>𡒊垢<EFBFBD>澆捆: { P, I, C, O, S } <EFBFBD>?{ population, intervention, comparison, outcome, studyDesign }
+  // 后端兼容: { P, I, C, O, S } 或 { population, intervention, comparison, outcome, studyDesign }
   // screeningService.ts 中有字段映射逻辑
   
-  // 蝑偦<EFBFBD>㗇<EFBFBD><EFBFBD>?  inclusionCriteria String   @map("inclusion_criteria") @db.Text
+  // 筛选标准
+  inclusionCriteria String   @map("inclusion_criteria") @db.Text
   exclusionCriteria String   @map("exclusion_criteria") @db.Text
   
-  // <EFBFBD>嗆<EFBFBD>?  status            String   @default("draft")
-  // <20>舫<EFBFBD>匧<EFBFBD>? draft, screening, completed
+  // 状态
+  status            String   @default("draft")
+  // 可选值: draft, screening, completed
   
-  // 蝑偦<EFBFBD>厰<EFBFBD>蝵?  screeningConfig   Json?    @map("screening_config")
+  // 筛选配置
+  screeningConfig   Json?    @map("screening_config")
   // 结构: { models: ["DeepSeek-V3", "Qwen-Max"], style: "standard" }
-  // <EFBFBD>𩤃<EFBFBD> 璅∪<E79285><E288AA>滨妍<E6BBA8>惩<EFBFBD>嚗?  // <20>滨垢撅閧內<E996A7>? DeepSeek-V3 <20>?API<50>? deepseek-chat
-  // <EFBFBD>滨垢撅閧內<EFBFBD>? Qwen-Max <20>?API<50>? qwen-max
+  // ⚠️ 模型名称映射：
+  // 前端展示名: DeepSeek-V3 → API名: deepseek-chat
+  // 前端展示名: Qwen-Max → API名: qwen-max
   // screeningService.ts 中有模型名映射逻辑
   
   // 关联
@@ -85,7 +98,7 @@ model AslScreeningProject {
 }
 ```
 
-**SQL銵函<EFBFBD><EFBFBD>?*:
+**SQL表结构**:
 ```sql
 CREATE TABLE asl_schema.screening_projects (
   id TEXT PRIMARY KEY,
@@ -108,12 +121,16 @@ CREATE INDEX idx_screening_projects_status ON asl_schema.screening_projects(stat
 
 ---
 
-### 2. <EFBFBD><EFBFBD>讃<EFBFBD>∠𤌍銵?(literatures) 潃?v3.0<EFBFBD>湔鰵
+### 2. 文献条目表 (literatures) ⭐ v3.0更新
 
-**Prisma璅∪<EFBFBD><EFBFBD>?*: `AslLiterature`  
+**Prisma模型名**: `AslLiterature`  
 **表名**: `asl_schema.literatures`
 
-**v3.0 <EFBFBD>湔鰵霂湔<EFBFBD>**嚗?- <20>?<3F>啣<EFBFBD> `stage` 摮埈挾嚗朞蕭頦芣<E9A0A6><E88AA3>桃<EFBFBD><E6A183>賢𪂹<E8B3A2><F0AA82B9><EFBFBD>imported <20>?title_screened <20>?pdf_acquired <20>?fulltext_screened <20>?data_extracted嚗?- <20>?<3F>啣<EFBFBD> PDF摮睃<E691AE>摮埈挾嚗𡁏𣈲<F0A1818F><F0A388B2>ify/OSS<53>屸<EFBFBD><E5B1B8><EFBFBD>嚗ǑpdfStorageType`, `pdfStorageRef`, `pdfStatus`嚗?- <20>?<3F>啣<EFBFBD> <20>冽<EFBFBD>摮睃<E691AE>摮埈挾嚗?*蝚血<E89D9A>鈭穃<E988AD><E7A983>蠘<EFBFBD><E8A098><EFBFBD><EFBFBD>摮睃<E691AE>撘閧鍂<E996A7>屸<EFBFBD><E5B1B8><EFBFBD>捆**嚗ǑfullTextStorageRef`, `fullTextUrl`嚗?- <20>?<3F>啣<EFBFBD>蝝Ｗ<E89D9D>嚗䫤stage`, `hasPdf`, `pdfStatus` <20>𣂼<EFBFBD><F0A382BC>亥砭<E4BAA5>扯<EFBFBD>
+**v3.0 更新说明**：
+- ✅ 新增 `stage` 字段：追踪文献生命周期（imported → title_screened → pdf_acquired → fulltext_screened → data_extracted）
+- ✅ 新增 PDF存储字段：支持Dify/OSS双适配（`pdfStorageType`, `pdfStorageRef`, `pdfStatus`）
+- ✅ 新增 全文存储字段：**符合云原生规范，存储引用而非内容**（`fullTextStorageRef`, `fullTextUrl`）
+- ✅ 新增索引：`stage`, `hasPdf`, `pdfStatus` 提升查询性能
 
 ```prisma
 model AslLiterature {
@@ -130,21 +147,25 @@ model AslLiterature {
   publicationYear   Int?     @map("publication_year")
   doi               String?
   
-  // 潃?v3.0 <EFBFBD>啣<EFBFBD>嚗𡁏<EFBFBD><EFBFBD>桅𧫴畾蛛<EFBFBD><EFBFBD>笔𦶢<EFBFBD>冽<EFBFBD>蝞∠<EFBFBD>嚗?  stage             String   @default("imported") @map("stage")
+  // ⭐ v3.0 新增：文献阶段（生命周期管理）
+  stage             String   @default("imported") @map("stage")
   // imported | title_screened | title_included | pdf_acquired | fulltext_screened | data_extracted
   
-  // 鈭穃<EFBFBD><EFBFBD>笔<EFBFBD><EFBFBD>典<EFBFBD>畾蛛<EFBFBD>V1.0 <20>嗆挾雿輻鍂嚗愢VP<56>嗆挾憸<E68CBE><E686B8>嚗?  pdfUrl            String?  @map("pdf_url")        // PDF霈輸䔮URL
+  // 云原生存储字段（V1.0 阶段使用，MVP阶段预留）
+  pdfUrl            String?  @map("pdf_url")        // PDF访问URL
   pdfOssKey         String?  @map("pdf_oss_key")    // OSS存储Key（用于删除）
   pdfFileSize       Int?     @map("pdf_file_size")  // 文件大小（字节）
   
-  // 潃?v3.0 <EFBFBD>啣<EFBFBD>嚗䥪DF摮睃<EFBFBD>嚗㇄ify/OSS<53>屸<EFBFBD><E5B1B8><EFBFBD>嚗?  hasPdf            Boolean  @default(false) @map("has_pdf")
+  // ⭐ v3.0 新增：PDF存储（Dify/OSS双适配）
+  hasPdf            Boolean  @default(false) @map("has_pdf")
   pdfStorageType    String?  @map("pdf_storage_type")  // "dify" | "oss"
   pdfStorageRef     String?  @map("pdf_storage_ref")   // Dify: document_id, OSS: object_key
   pdfStatus         String?  @map("pdf_status")        // "uploading" | "ready" | "failed"
   pdfUploadedAt     DateTime? @map("pdf_uploaded_at")
   
-  // 潃?v3.0 <EFBFBD>啣<EFBFBD>嚗𡁜<EFBFBD><EFBFBD><EFBFBD><EFBFBD>摰孵<EFBFBD><EFBFBD>剁<EFBFBD>鈭穃<EFBFBD><EFBFBD><EFBFBD><EFBFBD>摮睃<EFBFBD>撘閧鍂<EFBFBD>屸<EFBFBD><EFBFBD><EFBFBD>捆嚗?  fullTextStorageType String?  @map("full_text_storage_type")  // "dify" | "oss"
-  fullTextStorageRef  String?  @map("full_text_storage_ref")   // document_id <20>?object_key
+  // ⭐ v3.0 新增：全文内容存储（云原生：存储引用而非内容）
+  fullTextStorageType String?  @map("full_text_storage_type")  // "dify" | "oss"
+  fullTextStorageRef  String?  @map("full_text_storage_ref")   // document_id 或 object_key
   fullTextUrl         String?  @map("full_text_url")           // 访问URL
   fullTextFormat      String?  @map("full_text_format")        // "markdown" | "plaintext"
   fullTextSource      String?  @map("full_text_source")        // "nougat" | "pymupdf"
@@ -153,7 +174,7 @@ model AslLiterature {
   
   // 关联
   screeningResults  AslScreeningResult[]
-  fulltextScreeningResults AslFulltextScreeningResult[] // 潃?v3.0 <EFBFBD>啣<EFBFBD>
+  fulltextScreeningResults AslFulltextScreeningResult[] // ⭐ v3.0 新增
   
   createdAt         DateTime @default(now()) @map("created_at")
   updatedAt         DateTime @updatedAt @map("updated_at")
@@ -162,14 +183,14 @@ model AslLiterature {
   @@schema("asl_schema")
   @@index([projectId])
   @@index([doi])
-  @@index([stage])          // 潃?v3.0 <EFBFBD>啣<EFBFBD>
-  @@index([hasPdf])         // 潃?v3.0 <EFBFBD>啣<EFBFBD>
-  @@index([pdfStatus])      // 潃?v3.0 <EFBFBD>啣<EFBFBD>
+  @@index([stage])          // ⭐ v3.0 新增
+  @@index([hasPdf])         // ⭐ v3.0 新增
+  @@index([pdfStatus])      // ⭐ v3.0 新增
   @@unique([projectId, pmid])
 }
 ```
 
-**SQL銵函<EFBFBD><EFBFBD>?*嚗ǒ3.0嚗?
+**SQL表结构**（v3.0）:
 ```sql
 CREATE TABLE asl_schema.literatures (
   id TEXT PRIMARY KEY,
@@ -187,11 +208,13 @@ CREATE TABLE asl_schema.literatures (
   -- 文献阶段
   stage TEXT NOT NULL DEFAULT 'imported',
   
-  -- PDF摮睃<EFBFBD>嚗<EFBFBD>唂摮埈挾嚗㲋1.0憸<EFBFBD><EFBFBD>嚗?  pdf_url TEXT,
+  -- PDF存储（旧字段，V1.0预留）
+  pdf_url TEXT,
   pdf_oss_key TEXT,
   pdf_file_size INTEGER,
   
-  -- PDF摮睃<EFBFBD>嚗<EFBFBD>鰵摮埈挾嚗㷉ify/OSS<53>屸<EFBFBD><E5B1B8><EFBFBD>嚗?  has_pdf BOOLEAN NOT NULL DEFAULT false,
+  -- PDF存储（新字段，Dify/OSS双适配）
+  has_pdf BOOLEAN NOT NULL DEFAULT false,
   pdf_storage_type TEXT,
   pdf_storage_ref TEXT,
   pdf_status TEXT,
@@ -221,28 +244,32 @@ CREATE INDEX idx_literatures_has_pdf ON asl_schema.literatures(has_pdf);
 CREATE INDEX idx_literatures_pdf_status ON asl_schema.literatures(pdf_status);
 ```
 
-**摮埈挾霂湔<EFBFBD>**嚗?
+**字段说明**：
+
 | 字段 | 类型 | 说明 | 设计理由 |
 |------|------|------|----------|
-| `stage` | String | <EFBFBD><EFBFBD>讃<EFBFBD>嗆挾 | 餈質葵<E8B3AA><E891B5>讃<EFBFBD>冽㟲銝芣<E98A9D>蝔衤葉<E8A1A4><E89189><EFBFBD>蝵?|
-| `pdfStorageType` | String | PDF摮睃<EFBFBD>蝐餃<EFBFBD> | "dify"\|"oss"嚗峕𣈲<EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD>?|
+| `stage` | String | 文献阶段 | 追踪文献在整个流程中的位置 |
+| `pdfStorageType` | String | PDF存储类型 | "dify"\|"oss"，支持双适配器 |
 | `pdfStorageRef` | String | PDF存储引用 | Dify的document_id或OSS的object_key |
-| `fullTextStorageType` | String | <EFBFBD>冽<EFBFBD>摮睃<EFBFBD>蝐餃<EFBFBD> | 鈭穃<E988AD><E7A983><EFBFBD><EFBFBD>銝滨凒<E6BBA8>亙<EFBFBD><E4BA99>冽<EFBFBD>嚗<EFBFBD><E59A97>撘閧鍂 <20>?|
-| `fullTextStorageRef` | String | <EFBFBD>冽<EFBFBD>摮睃<EFBFBD>撘閧鍂 | <20><><EFBFBD>Dify<66>𤈛SS銝剔<E98A9D><E58994>冽<EFBFBD><E586BD><EFBFBD>﹝ <20>?|
+| `fullTextStorageType` | String | 全文存储类型 | 云原生：不直接存全文，存引用 ✅ |
+| `fullTextStorageRef` | String | 全文存储引用 | 指向Dify或OSS中的全文文档 ✅ |
 | `fullTextUrl` | String | 全文访问URL | 直接访问全文的URL |
 | `fullTextTokenCount` | Int | Token数量 | 用于成本估算和LLM调用优化 |
 
-**鈭穃<EFBFBD><EFBFBD>蠘挽霈∩漁<EFBFBD>?* 潃琜<E6BD83>
-- <20>?<3F>冽<EFBFBD><E586BD><EFBFBD>捆摮睃<E691AE><E79D83>汰SS/Dify嚗峕㺭<E5B395>桀<EFBFBD><E6A180>芸<EFBFBD>撘閧鍂嚗<E98D82>泵<EFBFBD><E6B3B5><EFBFBD><EFBFBD>毺<EFBFBD>閫<EFBFBD><E996AB>嚗?- <20>?<3F>舀<EFBFBD>Dify <20>?OSS<53>删<EFBFBD>餈<EFBFBD>宏嚗<E5AE8F>蘨<EFBFBD><E898A8><EFBFBD><EFBFBD>揢storageType嚗?- <20>?<3F>唳旿摨栞蝠<E6A09E>𧶏<EFBFBD><F0A7B68F>踹<EFBFBD>憭折<E686AD>TEXT摮埈挾
+**云原生设计亮点** ⭐：
+- ✅ 全文内容存储在OSS/Dify，数据库只存引用（符合云原生规范）
+- ✅ 支持Dify → OSS无缝迁移（只需切换storageType）
+- ✅ 数据库轻量，避免大量TEXT字段
 
 ---
 
 ### 3. 筛选结果表 (screening_results)
 
-**Prisma璅∪<EFBFBD><EFBFBD>?*: `AslScreeningResult`  
+**Prisma模型名**: `AslScreeningResult`  
 **表名**: `asl_schema.screening_results`
 
-**霈曇恣鈭桃<EFBFBD>**嚗𡁏𣈲<F0A1818F><F0A388B2><EFBFBD>璅∪<E79285>嚗㇄eepSeek + Qwen嚗匧僎銵屸<E98AB5>霂<EFBFBD><E99C82><EFBFBD><EFBFBD>鉄摰峕㟲<E5B395><E39FB2>ế<EFBFBD>准<EFBFBD><E58786><EFBFBD><EFBFBD>桀<EFBFBD><E6A180>脩<EFBFBD>璉<EFBFBD>瘚卝<E7989A>?
+**设计亮点**：支持双模型（DeepSeek + Qwen）并行验证，包含完整的判断、证据和冲突检测。
+
 ```prisma
 model AslScreeningResult {
   id                String   @id @default(uuid())
@@ -283,27 +310,36 @@ model AslScreeningResult {
   qwenSEvidence     String?  @map("qwen_s_evidence") @db.Text
   qwenReason        String?  @map("qwen_reason") @db.Text
   
-  // <EFBFBD>脩<EFBFBD><EFBFBD>嗆<EFBFBD>?  conflictStatus    String   @default("none") @map("conflict_status")
-  // <20>舫<EFBFBD>匧<EFBFBD>? none, conflict, resolved
+  // 冲突状态
+  conflictStatus    String   @default("none") @map("conflict_status")
+  // 可选值: none, conflict, resolved
   conflictFields    Json?    @map("conflict_fields")
   // 示例: ["P", "I", "conclusion"]
   
-  // <EFBFBD><EFBFBD>蝏<EFBFBD><EFBFBD>蝑吔<EFBFBD>Week 4 瘛瑕<E7989B><E79195>寞<EFBFBD>雿輻鍂嚗?  finalDecision     String?  @map("final_decision") // "include" | "exclude" | null
-  // 潃?Week 4 霂湔<E99C82>嚗帋犖撌亙<E6928C><E4BA99>詨<EFBFBD>霈曄蔭甇文<E79487>畾蛛<E795BE>雿靝蛹<E99D9D><E89BB9>蝏<EFBFBD><E89D8F>蝑?  // - include: 鈭箏極<E7AE8F>喳<EFBFBD>蝥喳<E89DA5>嚗<EFBFBD>虾<EFBFBD>賣綫蝧翠I撱箄悅嚗?  // - exclude: 鈭箏極<E7AE8F>喳<EFBFBD><E596B3>㘾膄嚗<E88684>虾<EFBFBD>賣綫蝧翠I撱箄悅嚗?  // - null: <20>芸<EFBFBD><E88AB8>賂<EFBFBD>雿輻鍂AI<41>喟<EFBFBD>
+  // 最终决策（Week 4 混合方案使用）
+  finalDecision     String?  @map("final_decision") // "include" | "exclude" | null
+  // ⭐ Week 4 说明：人工复核后设置此字段，作为最终决策
+  // - include: 人工决定纳入（可能推翻AI建议）
+  // - exclude: 人工决定排除（可能推翻AI建议）
+  // - null: 未复核，使用AI决策
   
   finalDecisionBy   String?  @map("final_decision_by") // userId
   finalDecisionAt   DateTime? @map("final_decision_at")
   
   exclusionReason   String?  @map("exclusion_reason") @db.Text
-  // 潃?Week 4 霂湔<EFBFBD>嚗帋犖撌亙‵<EFBFBD>嗵<EFBFBD><EFBFBD>㘾膄<EFBFBD>笔<EFBFBD>嚗<EFBFBD><EFBFBD><EFBFBD><EFBFBD>漣擃䀝<EFBFBD>AI<EFBFBD>𣂼<EFBFBD>嚗?  // - 憒<><E68692>finalDecision=exclude嚗峕迨摮埈挾摮睃<E691AE>鈭箏極憛怠<E6869B><E680A0><EFBFBD><EFBFBD><EFBFBD>?  // - 憒<><E68692>銝槃ull嚗<6C><E59A97>蝡航䌊<E888AA>其<EFBFBD>AI<41>斗鱏銝剜<E98A9D><E5899C>吔<EFBFBD>dsPJudgment/dsIJudgment蝑㚁<E89D91>
+  // ⭐ Week 4 说明：人工填写的排除原因（优先级高于AI提取）
+  // - 如果finalDecision=exclude，此字段存储人工填写的原因
+  // - 如果为null，前端自动从AI判断中提取（dsPJudgment/dsIJudgment等）
   // - Week 4 初筛结果页使用此字段显示排除原因
   
-  // AI憭<EFBFBD><EFBFBD><EFBFBD>嗆<EFBFBD>?  aiProcessingStatus String  @default("pending") @map("ai_processing_status")
-  // <20>舫<EFBFBD>匧<EFBFBD>? pending, processing, completed, failed
+  // AI处理状态
+  aiProcessingStatus String  @default("pending") @map("ai_processing_status")
+  // 可选值: pending, processing, completed, failed
   aiProcessedAt     DateTime? @map("ai_processed_at")
   aiErrorMessage    String?  @map("ai_error_message") @db.Text
   
-  // <EFBFBD>航蕭皞臭縑<EFBFBD>?  promptVersion     String   @default("v1.0.0") @map("prompt_version")
+  // 可追溯信息
+  promptVersion     String   @default("v1.0.0") @map("prompt_version")
   rawOutput         Json?    @map("raw_output") // 原始LLM输出（备份）
   
   createdAt         DateTime @default(now()) @map("created_at")
@@ -315,10 +351,11 @@ model AslScreeningResult {
   @@index([literatureId])
   @@index([conflictStatus])
   @@index([finalDecision])
-  @@unique([projectId, literatureId])  // 銝<EFBFBD>蝭<EFBFBD><EFBFBD><EFBFBD>桀銁銝<EFBFBD>銝芷★<EFBFBD>桐葉<EFBFBD>芣<EFBFBD>銝<EFBFBD>銝芰<EFBFBD><EFBFBD>厩<EFBFBD><EFBFBD>?}
+  @@unique([projectId, literatureId])  // 一篇文献在一个项目中只有一个筛选结果
+}
 ```
 
-**SQL銵函<EFBFBD><EFBFBD>?*嚗<><E59A97><EFBFBD>𣇉<EFBFBD>嚗?
+**SQL表结构**（简化版）:
 ```sql
 CREATE TABLE asl_schema.screening_results (
   id TEXT PRIMARY KEY,
@@ -353,19 +390,23 @@ CREATE TABLE asl_schema.screening_results (
   qwen_s_evidence TEXT,
   qwen_reason TEXT,
   
-  -- <EFBFBD>脩<EFBFBD><EFBFBD>嗆<EFBFBD>?  conflict_status TEXT NOT NULL DEFAULT 'none',
+  -- 冲突状态
+  conflict_status TEXT NOT NULL DEFAULT 'none',
   conflict_fields JSONB,
   
-  -- <EFBFBD><EFBFBD>蝏<EFBFBD><EFBFBD>蝑?  final_decision TEXT,
+  -- 最终决策
+  final_decision TEXT,
   final_decision_by TEXT,
   final_decision_at TIMESTAMP(3),
   exclusion_reason TEXT,
   
-  -- AI憭<EFBFBD><EFBFBD><EFBFBD>嗆<EFBFBD>?  ai_processing_status TEXT NOT NULL DEFAULT 'pending',
+  -- AI处理状态
+  ai_processing_status TEXT NOT NULL DEFAULT 'pending',
   ai_processed_at TIMESTAMP(3),
   ai_error_message TEXT,
   
-  -- <EFBFBD>航蕭皞臭縑<EFBFBD>?  prompt_version TEXT NOT NULL DEFAULT 'v1.0.0',
+  -- 可追溯信息
+  prompt_version TEXT NOT NULL DEFAULT 'v1.0.0',
   raw_output JSONB,
   
   created_at TIMESTAMP(3) NOT NULL DEFAULT CURRENT_TIMESTAMP,
@@ -388,7 +429,7 @@ CREATE INDEX idx_screening_results_final_decision ON asl_schema.screening_result
 
 ### 4. 筛选任务表 (screening_tasks)
 
-**Prisma璅∪<EFBFBD><EFBFBD>?*: `AslScreeningTask`  
+**Prisma模型名**: `AslScreeningTask`  
 **表名**: `asl_schema.screening_tasks`
 
 ```prisma
@@ -399,7 +440,7 @@ model AslScreeningTask {
   
   taskType          String   @map("task_type") // "title_abstract" | "full_text"
   status            String   @default("pending")
-  // <EFBFBD>舫<EFBFBD>匧<EFBFBD>? pending, running, completed, failed
+  // 可选值: pending, running, completed, failed
   
   // 进度统计
   totalItems        Int      @map("total_items")
@@ -426,7 +467,7 @@ model AslScreeningTask {
 }
 ```
 
-**SQL銵函<EFBFBD><EFBFBD>?*:
+**SQL表结构**:
 ```sql
 CREATE TABLE asl_schema.screening_tasks (
   id TEXT PRIMARY KEY,
@@ -454,12 +495,13 @@ CREATE INDEX idx_screening_tasks_status ON asl_schema.screening_tasks(status);
 
 ---
 
-### 5. <EFBFBD>冽<EFBFBD>憭滨<EFBFBD>隞餃𦛚銵?(fulltext_screening_tasks) 潃?v3.0<EFBFBD>啣<EFBFBD>
+### 5. 全文复筛任务表 (fulltext_screening_tasks) ⭐ v3.0新增
 
-**Prisma璅∪<EFBFBD><EFBFBD>?*: `AslFulltextScreeningTask`  
+**Prisma模型名**: `AslFulltextScreeningTask`  
 **表名**: `asl_schema.fulltext_screening_tasks`
 
-**霈曇恣<EFBFBD>格<EFBFBD>**嚗𡁶恣<F0A181B6><E681A3><EFBFBD><EFBFBD><EFBFBD><EFBFBD>蝑𤤿<E89D91><F0A4A4BF>孵<EFBFBD><E5ADB5><EFBFBD>遙<EFBFBD>∴<EFBFBD><E288B4>舀<EFBFBD><E88880>峕芋<E5B395>见僎銵諹<E98AB5><E8ABB9>具<EFBFBD><E585B7><EFBFBD><EFBFBD>祈蕭頦芥<E9A0A6><E88AA5><EFBFBD>蝥扳芋撘?
+**设计目标**：管理全文复筛的批处理任务，支持双模型并行调用、成本追踪、降级模式
+
 ```prisma
 model AslFulltextScreeningTask {
   id                String   @id @default(uuid())
@@ -471,7 +513,8 @@ model AslFulltextScreeningTask {
   modelB            String   @map("model_b")          // "qwen-max"
   promptVersion     String   @default("v1.0.0") @map("prompt_version")
   
-  // 隞餃𦛚<EFBFBD>嗆<EFBFBD>?  status            String   @default("pending")
+  // 任务状态
+  status            String   @default("pending")
   // "pending" | "running" | "completed" | "failed" | "cancelled"
   
   // 进度统计
@@ -479,7 +522,8 @@ model AslFulltextScreeningTask {
   processedCount    Int      @default(0) @map("processed_count")
   successCount      Int      @default(0) @map("success_count")
   failedCount       Int      @default(0) @map("failed_count")
-  degradedCount     Int      @default(0) @map("degraded_count")  // <EFBFBD>閙芋<EFBFBD>𧢲<EFBFBD><EFBFBD>?  
+  degradedCount     Int      @default(0) @map("degraded_count")  // 单模型成功
+  
   // 成本统计
   totalTokens       Int      @default(0) @map("total_tokens")
   totalCost         Float    @default(0) @map("total_cost")
@@ -507,7 +551,7 @@ model AslFulltextScreeningTask {
 }
 ```
 
-**SQL銵函<EFBFBD><EFBFBD>?*:
+**SQL表结构**:
 ```sql
 CREATE TABLE asl_schema.fulltext_screening_tasks (
   id TEXT PRIMARY KEY,
@@ -518,7 +562,8 @@ CREATE TABLE asl_schema.fulltext_screening_tasks (
   model_b TEXT NOT NULL,
   prompt_version TEXT NOT NULL DEFAULT 'v1.0.0',
   
-  -- 隞餃𦛚<EFBFBD>嗆<EFBFBD>?  status TEXT NOT NULL DEFAULT 'pending',
+  -- 任务状态
+  status TEXT NOT NULL DEFAULT 'pending',
   
   -- 进度统计
   total_count INTEGER NOT NULL,
@@ -552,25 +597,32 @@ CREATE INDEX idx_fulltext_screening_tasks_status ON asl_schema.fulltext_screenin
 CREATE INDEX idx_fulltext_screening_tasks_created_at ON asl_schema.fulltext_screening_tasks(created_at);
 ```
 
-**摮埈挾霂湔<EFBFBD>**嚗?
+**字段说明**：
+
 | 字段 | 类型 | 说明 |
 |------|------|------|
-| `modelA / modelB` | String | <EFBFBD>峕芋<EFBFBD>见<EFBFBD>蝘堆<EFBFBD>deepseek-v3 + qwen-max嚗?|
-| `degradedCount` | Int | <EFBFBD>閙芋<EFBFBD>𧢲<EFBFBD><EFBFBD>毺<EFBFBD>隞餃𦛚<EFBFBD>堆<EFBFBD>摰寥<EFBFBD><EFBFBD>箏<EFBFBD>嚗?|
-| `totalTokens` | Int | 蝝航恣Token雿輻鍂<EFBFBD>?|
-| `totalCost` | Float | 蝝航恣<EFBFBD>鞉𧋦嚗<EFBFBD><EFBFBD>嚗?|
-| `promptVersion` | String | Prompt<EFBFBD><EFBFBD>𧋦嚗<EFBFBD>虾餈賣滲嚗?|
+| `modelA / modelB` | String | 双模型名称（deepseek-v3 + qwen-max） |
+| `degradedCount` | Int | 单模型成功的任务数（容错机制） |
+| `totalTokens` | Int | 累计Token使用量 |
+| `totalCost` | Float | 累计成本（元） |
+| `promptVersion` | String | Prompt版本（可追溯） |
 
 ---
 
-### 6. <EFBFBD>冽<EFBFBD>憭滨<EFBFBD>蝏𤘪<EFBFBD>銵?(fulltext_screening_results) 潃?v3.0<EFBFBD>啣<EFBFBD>
+### 6. 全文复筛结果表 (fulltext_screening_results) ⭐ v3.0新增
 
-**Prisma璅∪<EFBFBD><EFBFBD>?*: `AslFulltextScreeningResult`  
+**Prisma模型名**: `AslFulltextScreeningResult`  
 **表名**: `asl_schema.fulltext_screening_results`
 
-**霈曇恣<EFBFBD>格<EFBFBD>**嚗𡁜<E59A97><F0A1819C>?2摮埈挾霂衣<E99C82>霂<EFBFBD>摯蝏𤘪<E89D8F>嚗峕𣈲<E5B395><F0A388B2><EFBFBD>璅∪<E79285>撖寞<E69296><E5AF9E><EFBFBD><EFBFBD>霂<EFBFBD><E99C82><EFBFBD>栶<EFBFBD><E6A0B6><EFBFBD>蝒<EFBFBD><E89D92>瘚?
-**霈曇恣鈭桃<E988AD>**嚗?- <20>?摰峕㟲<E5B395><E39FB2><EFBFBD>璅∪<E79285>蝏𤘪<E89D8F>嚗Êields + overall + logs嚗?- <20>?<3F>餃郎<E9A483>餉<EFBFBD>撉諹<E69289><E8ABB9>諹<EFBFBD><E8ABB9>桅曎撉諹<E69289>蝏𤘪<E89D8F>
-- <20>?<3F>脩<EFBFBD>璉<EFBFBD>瘚见<E7989A>憭齿瓲隡睃<E99AA1>蝥?- <20>?<3F>滨漣璅∪<E79285><E288AA>舀<EFBFBD>嚗<EFBFBD><E59A97>璅∪<E79285><E288AA>𣂼<EFBFBD>嚗?- <20>?JSON摮睃<E691AE>12摮埈挾霂<E68CBE>摯嚗<E691AF>泵<EFBFBD><E6B3B5><EFBFBD><EFBFBD>毺<EFBFBD>閫<EFBFBD><E996AB>嚗?
+**设计目标**：存储12字段详细评估结果，支持双模型对比、验证结果、冲突检测
+
+**设计亮点**：
+- ✅ 完整的双模型结果（fields + overall + logs）
+- ✅ 医学逻辑验证和证据链验证结果
+- ✅ 冲突检测和复核优先级
+- ✅ 降级模式支持（单模型成功）
+- ✅ JSON存储12字段评估（符合云原生规范）
+
 ```prisma
 model AslFulltextScreeningResult {
   id                String   @id @default(uuid())
@@ -581,7 +633,7 @@ model AslFulltextScreeningResult {
   literatureId      String   @map("literature_id")
   literature        AslLiterature @relation(fields: [literatureId], references: [id], onDelete: Cascade)
   
-  // ====== 璅∪<EFBFBD>A蝏𤘪<EFBFBD>嚗㇄eepSeek-V3嚗?=====
+  // ====== 模型A结果（DeepSeek-V3）======
   modelAName        String   @map("model_a_name")
   modelAStatus      String   @map("model_a_status")      // "success" | "failed"
   modelAFields      Json     @map("model_a_fields")      // 12字段评估 { field1: {...}, field2: {...}, ... }
@@ -592,7 +644,7 @@ model AslFulltextScreeningResult {
   modelACost        Float?   @map("model_a_cost")
   modelAError       String?  @map("model_a_error") @db.Text
   
-  // ====== 璅∪<EFBFBD>B蝏𤘪<EFBFBD>嚗㇋wen-Max嚗?=====
+  // ====== 模型B结果（Qwen-Max）======
   modelBName        String   @map("model_b_name")
   modelBStatus      String   @map("model_b_status")
   modelBFields      Json     @map("model_b_fields")
@@ -607,21 +659,22 @@ model AslFulltextScreeningResult {
   medicalLogicIssues Json?   @map("medical_logic_issues")  // MedicalLogicValidator输出
   evidenceChainIssues Json?  @map("evidence_chain_issues") // EvidenceChainValidator输出
   
-  // ====== <EFBFBD>脩<EFBFBD>璉<EFBFBD>瘚?======
+  // ====== 冲突检测 ======
   isConflict        Boolean  @default(false) @map("is_conflict")
   conflictSeverity  String?  @map("conflict_severity")        // "high" | "medium" | "low"
   conflictFields    String[] @map("conflict_fields")          // ["field1", "field9", "overall"]
   conflictDetails   Json?    @map("conflict_details")
-  reviewPriority    Int?     @map("review_priority")          // 0-100憭齿瓲隡睃<EFBFBD>蝥?  reviewDeadline    DateTime? @map("review_deadline")
+  reviewPriority    Int?     @map("review_priority")          // 0-100复核优先级
+  reviewDeadline    DateTime? @map("review_deadline")
   
-  // ====== <EFBFBD><EFBFBD>蝏<EFBFBD><EFBFBD>蝑?======
+  // ====== 最终决策 ======
   finalDecision     String?  @map("final_decision")           // "include" | "exclude" | null
   finalDecisionBy   String?  @map("final_decision_by")
   finalDecisionAt   DateTime? @map("final_decision_at")
   exclusionReason   String?  @map("exclusion_reason") @db.Text
   reviewNotes       String?  @map("review_notes") @db.Text
   
-  // ====== 憭<EFBFBD><EFBFBD><EFBFBD>嗆<EFBFBD>?======
+  // ====== 处理状态 ======
   processingStatus  String   @default("pending") @map("processing_status")
   // "pending" | "processing" | "completed" | "failed" | "degraded"
   isDegraded        Boolean  @default(false) @map("is_degraded")
@@ -629,7 +682,7 @@ model AslFulltextScreeningResult {
   
   processedAt       DateTime? @map("processed_at")
   
-  // ====== <EFBFBD>航蕭皞臭縑<EFBFBD>?======
+  // ====== 可追溯信息 ======
   promptVersion     String   @default("v1.0.0") @map("prompt_version")
   rawOutputA        Json?    @map("raw_output_a")
   rawOutputB        Json?    @map("raw_output_b")
@@ -645,10 +698,11 @@ model AslFulltextScreeningResult {
   @@index([isConflict])
   @@index([finalDecision])
   @@index([reviewPriority])
-  @@unique([projectId, literatureId])  // 銝<EFBFBD>蝭<EFBFBD><EFBFBD><EFBFBD>桀蘨<EFBFBD>劐<EFBFBD>銝芸<EFBFBD><EFBFBD><EFBFBD><EFBFBD>蝑𤤿<EFBFBD><EFBFBD>?}
+  @@unique([projectId, literatureId])  // 一篇文献只有一个全文复筛结果
+}
 ```
 
-**SQL銵函<EFBFBD><EFBFBD>?*嚗<><E59A97><EFBFBD>𣇉<EFBFBD>嚗<EFBFBD><E59A97><EFBFBD><EFBFBD><EFBFBD><EFBFBD>急<EFBFBD><E680A5>匧<EFBFBD>畾蛛<E795BE>:
+**SQL表结构**（简化版，实际包含所有字段）:
 ```sql
 CREATE TABLE asl_schema.fulltext_screening_results (
   id TEXT PRIMARY KEY,
@@ -682,25 +736,29 @@ CREATE TABLE asl_schema.fulltext_screening_results (
   medical_logic_issues JSONB,
   evidence_chain_issues JSONB,
   
-  -- <EFBFBD>脩<EFBFBD>璉<EFBFBD>瘚?  is_conflict BOOLEAN NOT NULL DEFAULT false,
+  -- 冲突检测
+  is_conflict BOOLEAN NOT NULL DEFAULT false,
   conflict_severity TEXT,
   conflict_fields TEXT[],
   conflict_details JSONB,
   review_priority INTEGER,
   review_deadline TIMESTAMP(3),
   
-  -- <EFBFBD><EFBFBD>蝏<EFBFBD><EFBFBD>蝑?  final_decision TEXT,
+  -- 最终决策
+  final_decision TEXT,
   final_decision_by TEXT,
   final_decision_at TIMESTAMP(3),
   exclusion_reason TEXT,
   review_notes TEXT,
   
-  -- 憭<EFBFBD><EFBFBD><EFBFBD>嗆<EFBFBD>?  processing_status TEXT NOT NULL DEFAULT 'pending',
+  -- 处理状态
+  processing_status TEXT NOT NULL DEFAULT 'pending',
   is_degraded BOOLEAN NOT NULL DEFAULT false,
   degraded_model TEXT,
   processed_at TIMESTAMP(3),
   
-  -- <EFBFBD>航蕭皞臭縑<EFBFBD>?  prompt_version TEXT NOT NULL DEFAULT 'v1.0.0',
+  -- 可追溯信息
+  prompt_version TEXT NOT NULL DEFAULT 'v1.0.0',
   raw_output_a JSONB,
   raw_output_b JSONB,
   
@@ -724,7 +782,8 @@ CREATE INDEX idx_fulltext_screening_results_final_decision ON asl_schema.fulltex
 CREATE INDEX idx_fulltext_screening_results_review_priority ON asl_schema.fulltext_screening_results(review_priority);
 ```
 
-**JSON摮埈挾蝷箔<EFBFBD>**嚗?
+**JSON字段示例**：
+
 **modelAFields (12字段评估)**:
 ```json
 {
@@ -732,7 +791,7 @@ CREATE INDEX idx_fulltext_screening_results_review_priority ON asl_schema.fullte
     "present": true,
     "completeness": "完整",
     "extractable": true,
-    "quote": "蝚砌<E89D9A>雿𡏭<E99BBF><F0A18FAD><EFBFBD>Zhang et al., <EFBFBD>𤏸”鈭?JAMA 2023...",
+    "quote": "第一作者：Zhang et al., 发表于 JAMA 2023...",
     "location": "Title page, Methods section",
     "note": "文献来源信息完整"
   },
@@ -747,9 +806,9 @@ CREATE INDEX idx_fulltext_screening_results_review_priority ON asl_schema.fullte
   "decision": "include",
   "confidence": 0.92,
   "keyIssues": [
-    "<22>𤩺㦤<F0A4A9BA>𡝗䲮瘜閙<E7989C>餈啣<E9A488><E595A3>?,
+    "随机化方法描述完整",
     "盲法实施清晰",
-    "蝏枏<E89D8F><E69E8F><EFBFBD><EFBFBD><EFBFBD>舀<EFBFBD><E88880>?
+    "结局指标可提取"
   ]
 }
 ```
@@ -767,7 +826,7 @@ CREATE INDEX idx_fulltext_screening_results_review_priority ON asl_schema.fullte
 {
   "field9": {
     "modelA": "完整",
-    "modelB": "銝滚<E98A9D><E6BB9A>?,
+    "modelB": "不完整",
     "severity": "high"
   }
 }
@@ -775,7 +834,8 @@ CREATE INDEX idx_fulltext_screening_results_review_priority ON asl_schema.fullte
 
 ---
 
-## <EFBFBD><EFBFBD> <20>唳旿<E594B3>喟頂<E5969F>橘<EFBFBD>v3.0<EFBFBD>湔鰵嚗?
+## 📊 数据关系图（v3.0更新）
+
 ```
 literature_screening_projects (1) ──< (N) literature_items
 literature_screening_projects (1) ──< (N) title_abstract_screening_results
@@ -785,41 +845,43 @@ literature_screening_projects (1) ──< (N) screening_tasks
 
 ---
 
-## <EFBFBD><EFBFBD> 蝝Ｗ<E89D9D>霈曇恣瘙<E681A3><E79899>鳴<EFBFBD>v3.0<EFBFBD>湔鰵嚗?
+## 🔍 索引设计汇总（v3.0更新）
+
 | 表名 | 索引字段 | 索引类型 | 说明 |
 |------|---------|---------|------|
 | screening_projects | user_id | B-tree | 用户项目查询 |
-| screening_projects | status | B-tree | <EFBFBD>嗆<EFBFBD><EFBFBD><EFBFBD><EFBFBD>?|
+| screening_projects | status | B-tree | 状态筛选 |
 | literatures | project_id | B-tree | 项目文献查询 |
 | literatures | doi | B-tree | DOI查重 |
-| literatures | stage 潃?| B-tree | <EFBFBD><EFBFBD>讃<EFBFBD>嗆挾<EFBFBD>亥砭 v3.0 |
-| literatures | has_pdf 潃?| B-tree | PDF<EFBFBD>瑕<EFBFBD><EFBFBD>嗆<EFBFBD>?v3.0 |
-| literatures | pdf_status 潃?| B-tree | PDF銝𠹺<EFBFBD><EFBFBD>嗆<EFBFBD>?v3.0 |
+| literatures | stage ⭐ | B-tree | 文献阶段查询 v3.0 |
+| literatures | has_pdf ⭐ | B-tree | PDF获取状态 v3.0 |
+| literatures | pdf_status ⭐ | B-tree | PDF上传状态 v3.0 |
 | literatures | (project_id, pmid) | Unique | 防止重复导入 |
 | screening_results | project_id | B-tree | 项目结果查询 |
 | screening_results | literature_id | B-tree | 文献结果查询 |
-| screening_results | conflict_status | B-tree | <EFBFBD>脩<EFBFBD>蝑偦<EFBFBD>?|
-| screening_results | final_decision | B-tree | <EFBFBD>喟<EFBFBD>蝑偦<EFBFBD>?|
-| screening_results | (project_id, literature_id) | Unique | <EFBFBD>臭<EFBFBD><EFBFBD>抒漲<EFBFBD>?|
+| screening_results | conflict_status | B-tree | 冲突筛选 |
+| screening_results | final_decision | B-tree | 决策筛选 |
+| screening_results | (project_id, literature_id) | Unique | 唯一性约束 |
 | screening_tasks | project_id | B-tree | 项目任务查询 |
-| screening_tasks | status | B-tree | 隞餃𦛚<EFBFBD>嗆<EFBFBD><EFBFBD><EFBFBD><EFBFBD>?|
-| fulltext_screening_tasks 潃?| project_id | B-tree | <EFBFBD>冽<EFBFBD>隞餃𦛚<EFBFBD>亥砭 v3.0 |
-| fulltext_screening_tasks 潃?| status | B-tree | 隞餃𦛚<EFBFBD>嗆<EFBFBD><EFBFBD><EFBFBD><EFBFBD>?v3.0 |
-| fulltext_screening_tasks 潃?| created_at | B-tree | <EFBFBD>園𡢿<EFBFBD>鍦<EFBFBD> v3.0 |
-| fulltext_screening_results 潃?| task_id | B-tree | 隞餃𦛚蝏𤘪<EFBFBD><EFBFBD>亥砭 v3.0 |
-| fulltext_screening_results 潃?| project_id | B-tree | 憿寧𤌍蝏𤘪<EFBFBD><EFBFBD>亥砭 v3.0 |
-| fulltext_screening_results 潃?| literature_id | B-tree | <EFBFBD><EFBFBD>讃蝏𤘪<EFBFBD><EFBFBD>亥砭 v3.0 |
-| fulltext_screening_results 潃?| is_conflict | B-tree | <EFBFBD>脩<EFBFBD>蝑偦<EFBFBD>?v3.0 |
-| fulltext_screening_results 潃?| final_decision | B-tree | <EFBFBD>喟<EFBFBD>蝑偦<EFBFBD>?v3.0 |
-| fulltext_screening_results 潃?| review_priority | B-tree | 憭齿瓲隡睃<EFBFBD>蝥?v3.0 |
-| fulltext_screening_results 潃?| (project_id, literature_id) | Unique | <EFBFBD>臭<EFBFBD><EFBFBD>抒漲<EFBFBD>?v3.0 |
+| screening_tasks | status | B-tree | 任务状态筛选 |
+| fulltext_screening_tasks ⭐ | project_id | B-tree | 全文任务查询 v3.0 |
+| fulltext_screening_tasks ⭐ | status | B-tree | 任务状态筛选 v3.0 |
+| fulltext_screening_tasks ⭐ | created_at | B-tree | 时间排序 v3.0 |
+| fulltext_screening_results ⭐ | task_id | B-tree | 任务结果查询 v3.0 |
+| fulltext_screening_results ⭐ | project_id | B-tree | 项目结果查询 v3.0 |
+| fulltext_screening_results ⭐ | literature_id | B-tree | 文献结果查询 v3.0 |
+| fulltext_screening_results ⭐ | is_conflict | B-tree | 冲突筛选 v3.0 |
+| fulltext_screening_results ⭐ | final_decision | B-tree | 决策筛选 v3.0 |
+| fulltext_screening_results ⭐ | review_priority | B-tree | 复核优先级 v3.0 |
+| fulltext_screening_results ⭐ | (project_id, literature_id) | Unique | 唯一性约束 v3.0 |
 
 **索引总数**: 25个（v3.0新增13个）  
 **唯一约束**: 4个（v3.0新增1个）
 
-**v3.0蝝Ｗ<EFBFBD>隡睃<EFBFBD>霂湔<EFBFBD>**嚗?- <20>?`literatures.stage`: 敹恍<E695B9><E6818D>䰻霂Ｙ鸌摰𡁻𧫴畾萇<E795BE><E89087><EFBFBD>讃嚗<E8AE83><E59A97>"pdf_acquired"敺<><E695BA><EFBFBD><EFBFBD><EFBFBD>蝑𨥈<E89D91>
-- <20>?`fulltext_screening_results.review_priority`: 隡睃<E99AA1>鈭箏極憭齿瓲<E9BDBF>笔<EFBFBD><E7AC94>鍦<EFBFBD>
-- <20>?`fulltext_screening_tasks.created_at`: 隞餃𦛚<E9A483><F0A69B9A>蟮<EFBFBD>亥砭隡睃<E99AA1>
+**v3.0索引优化说明**：
+- ✅ `literatures.stage`: 快速查询特定阶段的文献（如"pdf_acquired"待全文复筛）
+- ✅ `fulltext_screening_results.review_priority`: 优化人工复核队列排序
+- ✅ `fulltext_screening_tasks.created_at`: 任务历史查询优化
 
 ---
 
@@ -828,15 +890,15 @@ literature_screening_projects (1) ──< (N) screening_tasks
 ### PICO标准 (picoCriteria JSON)
 ```json
 {
-  "population": "<22>𠉛弦鈭箇黎嚗<E9BB8E><E59A97>嚗?<3F>讠<EFBFBD>撠輻<E692A0><E8BCBB>𣂷犖<F0A382B7><E78A96><EFBFBD>?,
-  "intervention": "撟脤<E6929F><E884A4>芣鴌嚗<E9B48C><E59A97>嚗锭GLT2<54>穃<EFBFBD><E7A983>?,
-  "comparison": "撖寧<E69296>嚗<EFBFBD><E59A97>嚗𡁜<E59A97><F0A1819C>啣<EFBFBD><E595A3>硋虜閫<E8999C><E996AB>瘜?,
+  "population": "研究人群，如：2型糖尿病成人患者",
+  "intervention": "干预措施，如：SGLT2抑制剂",
+  "comparison": "对照，如：安慰剂或常规疗法",
   "outcome": "结局指标，如：心血管结局",
-  "studyDesign": "<22>𠉛弦霈曇恣嚗<E681A3><E59A97>嚗𡁻<E59A97><F0A181BB>箏笆<E7AE8F>扯<EFBFBD>撉?(RCT)"
+  "studyDesign": "研究设计，如：随机对照试验 (RCT)"
 }
 ```
 
-### 蝑偦<EFBFBD>厰<EFBFBD>蝵?(screeningConfig JSON)
+### 筛选配置 (screeningConfig JSON)
 ```json
 {
   "models": ["deepseek-chat", "qwen-max"],
@@ -863,62 +925,97 @@ literature_screening_projects (1) ──< (N) screening_tasks
 ## 🔒 数据安全
 
 ### Schema隔离
-- 雿輻鍂 `asl_schema` 銝𤾸<E98A9D>隞𡝗芋<F0A19D97>埈㺭<E59F88>桅<EFBFBD>蝳?- <20>冽<EFBFBD>銵典銁 `platform_schema`嚗𣬚<E59A97>銝<EFBFBD>蝞∠<E89D9E>
+- 使用 `asl_schema` 与其他模块数据隔离
+- 用户表在 `platform_schema`，统一管理
 
 ### 级联删除
-- <20>𣳇膄<F0A3B387>冽<EFBFBD> <20>?<3F>芸𢆡<E88AB8>𣳇膄<F0A3B387><E88684><EFBFBD>厩<EFBFBD><E58EA9>厰★<E58EB0>桀<EFBFBD><E6A180>唾<EFBFBD><E594BE>唳旿
-- <20>𣳇膄憿寧𤌍 <20>?<3F>芸𢆡<E88AB8>𣳇膄<F0A3B387><E88684>讃<EFBFBD><E8AE83><EFBFBD><EFBFBD>栶<EFBFBD><E6A0B6>遙<EFBFBD>?- <20>𣳇膄<F0A3B387><E88684>讃 <20>?<3F>芸𢆡<E88AB8>𣳇膄蝑偦<E89D91>厩<EFBFBD><E58EA9>?
-### <20>臭<EFBFBD><E887AD>抒漲<E68A92>?- <20>䔶<EFBFBD>憿寧𤌍銝致MID<49>臭<EFBFBD>嚗<EFBFBD><E59A97>霈豢<E99C88>PMID嚗?- <20>䔶<EFBFBD>憿寧𤌍銝凋<E98A9D>蝭<EFBFBD><E89DAD><EFBFBD>桀蘨<E6A180>劐<EFBFBD>銝芰<E98A9D><E88AB0>厩<EFBFBD><E58EA9>?
+- 删除用户 → 自动删除所有筛选项目及关联数据
+- 删除项目 → 自动删除文献、结果、任务
+- 删除文献 → 自动删除筛选结果
+
+### 唯一性约束
+- 同一项目中PMID唯一（允许无PMID）
+- 同一项目中一篇文献只有一个筛选结果
+
 ---
 
-## <EFBFBD><EFBFBD> <20>唳旿<E594B3>誯<EFBFBD>隡?
-| 憿寧𤌍閫<F0A48C8D>芋 | <20><>讃<EFBFBD>?| 蝑偦<E89D91>厩<EFBFBD><E58EA9>?| 摮睃<E691AE>蝛粹𡢿 |
+## 📈 数据量预估
+
+| 项目规模 | 文献数 | 筛选结果 | 存储空间 |
 |---------|--------|---------|----------|
 | 小型 | 100-500 | 100-500 | < 10 MB |
 | 中型 | 500-2000 | 500-2000 | 10-50 MB |
 | 大型 | 2000-5000 | 2000-5000 | 50-200 MB |
-| 頞<EFBFBD>之<EFBFBD>?| 5000+ | 5000+ | 200 MB+ |
+| 超大型 | 5000+ | 5000+ | 200 MB+ |
 
 **单条记录大小估算**:
 - 文献条目：~2-5 KB
-- 蝑偦<E89D91>厩<EFBFBD><E58EA9>頣<EFBFBD>~5-10 KB嚗<42>鉄<EFBFBD>峕芋<E5B395>见ế<E8A781>剖<EFBFBD>霂<EFBFBD>旿嚗?
+- 筛选结果：~5-10 KB（含双模型判断和证据）
+
 ---
 
-## <EFBFBD>?<3F>𡒊賒閫<E8B392><E996AB>
+## ⏳ 后续规划
+
+### Phase 2 (全文复筛) ✅ v3.0已完成
+- [x] 扩展 `literatures` 表（生命周期管理）
+- [x] 添加 `fulltext_screening_tasks` 表
+- [x] 添加 `fulltext_screening_results` 表（12字段）
+
+### Phase 3 (数据提取) 待开发
+- [ ] 复用 `fulltext_screening_tasks` 表（切换模式）
+- [ ] 复用 `fulltext_screening_results` 表（存储提取数据）
+- [ ] 或新增 `data_extraction_results` 表（如需独立）
+
+### Phase 4 (质量评估) 待规划
+- [ ] 质量评估结果表
+- [ ] 偏倚风险评估表
+- [ ] GRADE证据质量表
 
-### Phase 2 (<28>冽<EFBFBD>憭滨<E686AD>) <20>?v3.0撌脣<EFBFBD><EFBFBD>?- [x] <20>拙<EFBFBD> `literatures` 銵剁<E98AB5><E58981>笔𦶢<E7AC94>冽<EFBFBD>蝞∠<E89D9E>嚗?- [x] 瘛餃<E7989B> `fulltext_screening_tasks` 銵?- [x] 瘛餃<E7989B> `fulltext_screening_results` 銵剁<E98AB5>12摮埈挾嚗?
-### Phase 3 (<28>唳旿<E594B3>𣂼<EFBFBD>) 敺<><E695BA><EFBFBD>?- [ ] 憭滨鍂 `fulltext_screening_tasks` 銵剁<E98AB5><E58981><EFBFBD>揢璅∪<E79285>嚗?- [ ] 憭滨鍂 `fulltext_screening_results` 銵剁<E98AB5>摮睃<E691AE><E79D83>𣂼<EFBFBD><F0A382BC>唳旿嚗?- [ ] <20>𡝗鰵憓?`data_extraction_results` 銵剁<E98AB5>憒<EFBFBD><E68692><EFBFBD>祉<EFBFBD>嚗?
-### Phase 4 (韐券<E99F90>霂<EFBFBD>摯) 敺<><E695BA><EFBFBD>?- [ ] 韐券<E99F90>霂<EFBFBD>摯蝏𤘪<E89D8F>銵?- [ ] <20>誩<EFBFBD>𡁻<EFBFBD><F0A181BB>抵<EFBFBD>隡啗”
-- [ ] GRADE霂<45>旿韐券<E99F90>銵?
 ---
 
 ## 📝 v3.0 设计决策记录
 
-### <EFBFBD>喟<EFBFBD>1: <20>冽<EFBFBD><E586BD><EFBFBD>捆摮睃<E691AE>撘閧鍂<E996A7>屸<EFBFBD><E5B1B8>湔𦻖摮睃<E691AE> <20>?
+### 决策1: 全文内容存储引用而非直接存储 ✅
+
 **问题**：全文内容是否存储在数据库？
 
-**<EFBFBD>寞<EFBFBD>撖寞<EFBFBD>**嚗?| <20>寞<EFBFBD> | 隡条<E99AA1> | 蝻箇<E89DBB> |
+**方案对比**：
+| 方案 | 优点 | 缺点 |
 |------|------|------|
-| 摮姯EXT | LLM靚<EFBFBD>鍂敹?| 餈肽<E9A488>鈭穃<E988AD><E7A983>蠘<EFBFBD><E8A098><EFBFBD><EFBFBD><EFBFBD>唳旿摨栞<E691A8><E6A09E>?|
-| 摮睃<EFBFBD><EFBFBD>?| 蝚血<E89D9A>閫<EFBFBD><E996AB>嚗諹蝠<E8ABB9>?| LLM靚<4D>鍂憓𧼮<E68693>100-200ms |
+| 存TEXT | LLM调用快 | 违背云原生规范，数据库臃肿 |
+| 存引用 | 符合规范，轻量 | LLM调用增加100-200ms |
+
+**决策**：✅ 采用方案2（存引用）
+- 符合云原生存储与计算分离原则
+- 支持超大文献（>1MB）
+- RDS存储成本是OSS的5-10倍
+
+### 决策2: 12字段使用JSON存储 ✅
+
+**问题**：12字段是拆分为列还是JSON存储？
 
-**<2A>喟<EFBFBD>**嚗尠<E59A97> <20><>鍂<EFBFBD>寞<EFBFBD>2嚗<32><E59A97>撘閧鍂嚗?- 蝚血<E89D9A>鈭穃<E988AD><E7A983>笔<EFBFBD><E7AC94>其<EFBFBD>霈∠<E99C88><E288A0><EFBFBD>氖<EFBFBD>笔<EFBFBD>
-- <20>舀<EFBFBD>頞<EFBFBD>之<EFBFBD><E4B98B>讃嚗?1MB嚗?- RDS摮睃<E691AE><E79D83>鞉𧋦<E99E89>烙SS<53>?-10<31>?
-### <20>喟<EFBFBD>2: 12摮埈挾雿輻鍂JSON摮睃<E691AE> <20>?
-**<2A>桅<EFBFBD>**嚗?2摮埈挾<E59F88>舀<EFBFBD><E88880><EFBFBD>蛹<EFBFBD>𡑒<EFBFBD><F0A19192>浥SON摮睃<E691AE>嚗?
 **决策**：✅ 使用PostgreSQL JSONB
-- 銝漤<E98A9D>閬<EFBFBD><E996AC><EFBFBD>祆䰻霂Ｘ<E99C82>銝芸<E98A9D>畾萄<E795BE><E89084>?- 摮埈挾蝏𤘪<E89D8F>憭齿<E686AD>嚗?銝芸<E98A9D>摮埈挾嚗?- JSONB<4E>扯<EFBFBD>隡条<E99AA1>銝娍𣈲<E5A88D><F0A388B2>IN蝝Ｗ<E89D9D>
+- 不需要单独查询某个字段内部
+- 字段结构复杂（6个子字段）
+- JSONB性能优秀且支持GIN索引
 
-### <EFBFBD>喟<EFBFBD>3: <20>祉<EFBFBD><E7A589>冽<EFBFBD>憭滨<E686AD>蝏𤘪<E89D8F>銵?<3F>?
-**<2A>桅<EFBFBD>**嚗𡁏糓<F0A1818F>血<EFBFBD><E8A180>?`screening_results` 銵剁<E98AB5>
+### 决策3: 独立全文复筛结果表 ✅
+
+**问题**：是否复用 `screening_results` 表？
+
+**决策**：✅ 新增独立表 `fulltext_screening_results`
+- 数据结构完全不同（PICOS vs 12字段）
+- 避免字段冗余和逻辑耦合
+- 便于独立维护和优化
 
-**<2A>喟<EFBFBD>**嚗尠<E59A97> <20>啣<EFBFBD><E595A3>祉<EFBFBD>銵?`fulltext_screening_results`
-- <20>唳旿蝏𤘪<E89D8F>摰<EFBFBD><E691B0>銝滚<E98A9D>嚗㇊ICOS vs 12摮埈挾嚗?- <20>踹<EFBFBD>摮埈挾<E59F88>𦯀<EFBFBD><F0A6AF80>屸<EFBFBD>餉<EFBFBD><E9A489>血<EFBFBD>
-- 靘蹂<E99D98><E8B982>祉<EFBFBD>蝏湔擪<E6B994>䔶<EFBFBD><E494B6>?
 ---
 
-**<EFBFBD><EFBFBD>﹝<EFBFBD><EFBFBD>𧋦嚗?* v3.0  
-**<EFBFBD><EFBFBD><EFBFBD>擧凒<EFBFBD>堆<EFBFBD>** 2025-11-22嚗㇄ay 4嚗𡁜<EFBFBD><EFBFBD><EFBFBD><EFBFBD>蝑𥟇㺭<EFBFBD>桀<EFBFBD>霈曇恣嚗? 
-**蝏湔擪<EFBFBD><EFBFBD><EFBFBD>** AI<EFBFBD>箄<EFBFBD><EFBFBD><EFBFBD>讃撘<EFBFBD><EFBFBD>穃𣪧<EFBFBD>?
-**<EFBFBD><EFBFBD>𧋦<EFBFBD><EFBFBD>蟮**嚗?- v3.0 (2025-11-22): <20>冽<EFBFBD>憭滨<E686AD><E6BBA8>唳旿摨栞挽霈∴<E99C88><E288B4>啣<EFBFBD>3銝芾”<E88ABE>𣬚㮾<F0A3AC9A>喳<EFBFBD>畾?- v2.2 (2025-11-21): Week 4蝏蠘恣<E8A098>蠘<EFBFBD>摰峕<E691B0>
-- v2.0 (2025-11-18): <20><><EFBFBD><EFBFBD>萘<EFBFBD><E89098>唳旿摨栞挽霈?- v1.0 (2025-10-29): <20>嘥<EFBFBD><E598A5><EFBFBD>𧋦
+**文档版本：** v3.0  
+**最后更新：** 2025-11-22（Day 4：全文复筛数据库设计）  
+**维护者：** AI智能文献开发团队
+
+**版本历史**：
+- v3.0 (2025-11-22): 全文复筛数据库设计，新增3个表和相关字段
+- v2.2 (2025-11-21): Week 4统计功能完成
+- v2.0 (2025-11-18): 标题初筛数据库设计
+- v1.0 (2025-10-29): 初始版本