feat(rvw): Complete RVW module development Phase 1-3

Summary:
- Migrate backend to modules/rvw with v2 API routes (/api/v2/rvw)
- Add new database fields: selectedAgents, editorialScore, methodologyStatus, picoExtract, isArchived
- Create frontend module in frontend-v2/src/modules/rvw
- Implement Dashboard with task list, filtering, batch operations
- Implement ReportDetail with dual tabs (editorial/methodology)
- Implement AgentModal for intelligent agent selection
- Register RVW module in moduleRegistry.ts
- Add navigation entry in TopNavigation
- Update documentation for RVW module status (v3.0)
- Update system status document (v2.9)

Features:
- User can select agents: editorial, methodology, or both
- Support batch task execution
- Task status filtering
- Replace console.log with logger service
- Maintain v1 API backward compatibility

Tested: Frontend and backend verified locally
Status: 85% complete (Phase 1-3 done)

backend/src/modules/rvw/services/utils.ts

@@ -0,0 +1,116 @@
+/**
+ * RVW稿件审查模块 - 工具函数
+ * @module rvw/services/utils
+ */
+
+import { MethodologyReview, MethodologyStatus } from '../types/index.js';
+
+/**
+ * 从LLM响应中解析JSON
+ * 支持多种格式：纯JSON、```json代码块、混合文本
+ */
+export function parseJSONFromLLMResponse<T>(content: string): T {
+  try {
+    // 1. 尝试直接解析
+    return JSON.parse(content) as T;
+  } catch {
+    // 2. 尝试提取```json代码块
+    const jsonMatch = content.match(/```json\s*\n?([\s\S]*?)\n?```/);
+    if (jsonMatch) {
+      try {
+        return JSON.parse(jsonMatch[1].trim()) as T;
+      } catch {
+        // 继续尝试其他方法
+      }
+    }
+
+    // 3. 尝试提取{}或[]包裹的内容
+    const objectMatch = content.match(/(\{[\s\S]*\})/);
+    if (objectMatch) {
+      try {
+        return JSON.parse(objectMatch[1]) as T;
+      } catch {
+        // 继续尝试其他方法
+      }
+    }
+
+    const arrayMatch = content.match(/(\[[\s\S]*\])/);
+    if (arrayMatch) {
+      try {
+        return JSON.parse(arrayMatch[1]) as T;
+      } catch {
+        // 失败
+      }
+    }
+
+    // 4. 所有尝试都失败
+    throw new Error('无法从LLM响应中解析JSON');
+  }
+}
+
+/**
+ * 根据方法学评估结果判断状态
+ * @param review 方法学评估结果
+ * @returns pass | warn | fail
+ */
+export function getMethodologyStatus(review: MethodologyReview | null | undefined): MethodologyStatus | undefined {
+  if (!review) return undefined;
+
+  const score = review.overall_score;
+
+  if (score >= 80) return 'pass';
+  if (score >= 60) return 'warn';
+  return 'fail';
+}
+
+/**
+ * 根据选择的智能体计算综合分数
+ * @param editorialScore 稿约规范性分数
+ * @param methodologyScore 方法学分数
+ * @param agents 选择的智能体
+ * @returns 综合分数
+ */
+export function calculateOverallScore(
+  editorialScore: number | null | undefined,
+  methodologyScore: number | null | undefined,
+  agents: string[]
+): number | null {
+  const hasEditorial = agents.includes('editorial') && editorialScore != null;
+  const hasMethodology = agents.includes('methodology') && methodologyScore != null;
+
+  if (hasEditorial && hasMethodology) {
+    // 两个都选：稿约40% + 方法学60%
+    return editorialScore! * 0.4 + methodologyScore! * 0.6;
+  } else if (hasEditorial) {
+    // 只选规范性
+    return editorialScore!;
+  } else if (hasMethodology) {
+    // 只选方法学
+    return methodologyScore!;
+  }
+
+  return null;
+}
+
+/**
+ * 验证智能体选择
+ * @param agents 选择的智能体列表
+ * @throws 如果选择无效
+ */
+export function validateAgentSelection(agents: string[]): void {
+  if (!agents || agents.length === 0) {
+    throw new Error('请至少选择一个智能体');
+  }
+
+  const validAgents = ['editorial', 'methodology'];
+  for (const agent of agents) {
+    if (!validAgents.includes(agent)) {
+      throw new Error(`无效的智能体类型: ${agent}`);
+    }
+  }
+
+  if (agents.length > 2) {
+    throw new Error('最多只能选择2个智能体');
+  }
+}
+