fix(rvw): speed up review flow and complete forensics export

Run RVW skills in controlled parallel mode and persist per-skill progress so users can view completed tabs during execution.
Also include data-forensics tables in Word export and refresh the RVW module status documentation.

Made-with: Cursor

backend/src/modules/rvw/controllers/reviewController.ts

@@ -315,6 +315,11 @@ export async function getTaskDetail(
     logger.info('[RVW:Controller] 获取任务详情', { taskId });
 
     const task = await reviewService.getTaskDetail(userId, taskId);
+    const contextData = task.contextData as {
+      forensicsResult?: unknown;
+      clinicalReview?: unknown;
+      skillProgress?: Record<string, unknown>;
+    } | null;
 
     // 🆕 直接使用新字段
     return reply.send({
@@ -336,6 +341,11 @@ export async function getTaskDetail(
         durationSeconds: task.durationSeconds,
         errorMessage: task.errorMessage,
         errorDetails: task.errorDetails ?? undefined,
+        editorialReview: task.editorialReview ?? undefined,
+        methodologyReview: task.methodologyReview ?? undefined,
+        forensicsResult: contextData?.forensicsResult ?? undefined,
+        clinicalReview: contextData?.clinicalReview ?? undefined,
+        reviewProgress: contextData?.skillProgress ?? undefined,
       },
     });
   } catch (error) {

backend/src/modules/rvw/skills/core/executor.ts

@@ -125,45 +125,49 @@ export class SkillExecutor<TContext extends BaseSkillContext = SkillContext> {
           taskId: context.taskId,
           skillIds: stage.map(s => s.skillId),
         });
+        const stageConcurrency = Math.max(1, profile.globalConfig?.maxConcurrency ?? stage.length);
+        for (let offset = 0; offset < stage.length; offset += stageConcurrency) {
+          const chunk = stage.slice(offset, offset + stageConcurrency);
+          const promises = chunk.map(item => this.executePipelineItem(item, context, profile));
+          const settled = await Promise.allSettled(promises);
 
-        const promises = stage.map(item => this.executePipelineItem(item, context, profile));
+          for (let i = 0; i < chunk.length; i++) {
-        const settled = await Promise.allSettled(promises);
+            const outcome = settled[i];
-
+            const currentItem = chunk[i];
-        for (let i = 0; i < stage.length; i++) {
+            if (outcome.status === 'fulfilled') {
-          const outcome = settled[i];
+              const result = outcome.value;
-          if (outcome.status === 'fulfilled') {
+              if (result) {
-            const result = outcome.value;
+                results.push(result);
-            if (result) {
+                context.previousResults.push(result);
-              results.push(result);
+                const skill = SkillRegistry.get(currentItem.skillId);
-              context.previousResults.push(result);
+                if (skill) this.updateContextWithResult(context, skill, result);
-              const skill = SkillRegistry.get(stage[i].skillId);
+              }
-              if (skill) this.updateContextWithResult(context, skill, result);
+            } else {
+              // Promise 本身 rejected — 极端情况下的兜底
+              const errorMessage = outcome.reason instanceof Error
+                ? outcome.reason.message
+                : String(outcome.reason);
+              logger.error('[SkillExecutor] Parallel skill promise rejected (uncaught)', {
+                skillId: currentItem.skillId,
+                taskId: context.taskId,
+                error: errorMessage,
+              });
+              const now = new Date();
+              results.push({
+                skillId: currentItem.skillId,
+                skillName: currentItem.skillId,
+                status: 'error',
+                issues: [{
+                  severity: 'ERROR',
+                  type: SkillErrorCodes.SKILL_EXECUTION_ERROR,
+                  message: `${currentItem.skillId} 执行异常: ${errorMessage}`,
+                }],
+                error: errorMessage,
+                executionTime: 0,
+                startedAt: now,
+                completedAt: now,
+              });
             }
-          } else {
-            // Promise 本身 rejected — 极端情况下的兜底
-            const errorMessage = outcome.reason instanceof Error
-              ? outcome.reason.message
-              : String(outcome.reason);
-            logger.error('[SkillExecutor] Parallel skill promise rejected (uncaught)', {
-              skillId: stage[i].skillId,
-              taskId: context.taskId,
-              error: errorMessage,
-            });
-            const now = new Date();
-            results.push({
-              skillId: stage[i].skillId,
-              skillName: stage[i].skillId,
-              status: 'error',
-              issues: [{
-                severity: 'ERROR',
-                type: SkillErrorCodes.SKILL_EXECUTION_ERROR,
-                message: `${stage[i].skillId} 执行异常: ${errorMessage}`,
-              }],
-              error: errorMessage,
-              executionTime: 0,
-              startedAt: now,
-              completedAt: now,
-            });
           }
         }
       }

backend/src/modules/rvw/skills/core/profile.ts

@@ -29,6 +29,7 @@ export const DEFAULT_PROFILE: JournalProfile = {
         tolerancePercent: 0.1,
       },
       timeout: 300000,  // 5min: Python + LLM核查(内部180s超时降级) + 长文档余量
+      parallelGroup: 'llm-review', // 与其余模块并行，缩短总时长
     },
     {
       skillId: 'EditorialSkill',
@@ -57,6 +58,7 @@ export const DEFAULT_PROFILE: JournalProfile = {
     strictness: 'STANDARD',
     continueOnError: true,
     timeoutMultiplier: 1.0,
+    maxConcurrency: 4, // 受控并行：默认同时跑4个模块
   },
 };
 

backend/src/modules/rvw/workers/reviewWorker.ts

@@ -34,6 +34,7 @@ import {
   createPartialContextFromTask,
   registerBuiltinSkills,
   ExecutionSummary,
+  SkillResult,
 } from '../skills/index.js';
 
 /**
@@ -307,6 +308,7 @@ export async function registerReviewWorker() {
             },
             forensicsResult: skillsSummary.results.find(r => r.skillId === 'DataForensicsSkill')?.data,
             clinicalReview: clinicalResult,
+            skillProgress: buildSkillProgressMap(skillsSummary.results),
           }
         : null;
 
@@ -439,6 +441,19 @@ function buildErrorDetails(summary: ExecutionSummary): Record<string, unknown> {
   };
 }
 
+function buildSkillProgressMap(results: SkillResult[]): Record<string, unknown> {
+  const progress: Record<string, unknown> = {};
+  for (const item of results) {
+    progress[item.skillId] = {
+      status: item.status,
+      executionTime: item.executionTime,
+      completedAt: item.completedAt.toISOString(),
+      error: item.error || null,
+    };
+  }
+  return progress;
+}
+
 /**
  * 使用 V2.0 Skills 架构执行审查
  */
@@ -480,9 +495,73 @@ async function executeWithSkills(
     fileSize,
   });
 
+  const currentTask = await prisma.reviewTask.findUnique({
+    where: { id: taskId },
+    select: { contextData: true },
+  });
+  const incrementalContext =
+    ((currentTask?.contextData as Record<string, unknown> | null) || {});
+  const runningContext: Record<string, unknown> = { ...incrementalContext };
+  const skillProgress =
+    ((incrementalContext.skillProgress as Record<string, unknown> | undefined) || {});
+  let persistQueue: Promise<void> = Promise.resolve();
+
+  const persistSkillResult = (result: SkillResult) => {
+    persistQueue = persistQueue.then(async () => {
+      skillProgress[result.skillId] = {
+        status: result.status,
+        executionTime: result.executionTime,
+        completedAt: result.completedAt.toISOString(),
+        error: result.error || null,
+      };
+
+      const nextContext: Record<string, unknown> = {
+        ...runningContext,
+        skillProgress: { ...skillProgress },
+      };
+
+      const updateData: Record<string, unknown> = {
+        contextData: nextContext as Prisma.InputJsonValue,
+      };
+
+      if (result.skillId === 'DataForensicsSkill' && result.data) {
+        nextContext.forensicsResult = result.data;
+      }
+      if (result.skillId === 'ClinicalAssessmentSkill' && result.data) {
+        nextContext.clinicalReview = result.data;
+      }
+      if (result.skillId === 'EditorialSkill' && result.data) {
+        const editorialData = result.data as EditorialReview;
+        updateData.editorialReview = editorialData as unknown as Prisma.InputJsonValue;
+        updateData.editorialScore = editorialData.overall_score ?? null;
+      }
+      if (result.skillId === 'MethodologySkill' && result.data) {
+        const methodologyData = result.data as MethodologyReview;
+        updateData.methodologyReview = methodologyData as unknown as Prisma.InputJsonValue;
+        updateData.methodologyScore = methodologyData.overall_score ?? null;
+        updateData.methodologyStatus = getMethodologyStatus(methodologyData);
+      }
+
+      updateData.contextData = nextContext as Prisma.InputJsonValue;
+      Object.assign(runningContext, nextContext);
+
+      await prisma.reviewTask.update({
+        where: { id: taskId },
+        data: updateData as Prisma.ReviewTaskUpdateInput,
+      });
+    });
+
+    return persistQueue;
+  };
+
   // 执行 Pipeline
-  const executor = new SkillExecutor();
+  const executor = new SkillExecutor({
+    onSkillComplete: async (_skillId, result) => {
+      await persistSkillResult(result);
+    },
+  });
   const summary = await executor.execute(profile, partialContext);
+  await persistQueue;
 
   // 输出执行结果
   console.log(`\n  📊 Skills 执行结果:`);

docs/03-业务模块/RVW-稿件审查系统/00-模块当前状态与开发指南.md

@@ -1,10 +1,10 @@
 # RVW稿件审查模块 - 当前状态与开发指南
 
-> **文档版本：** v6.0  
+> **文档版本：** v6.1  
 > **创建日期：** 2026-01-07  
-> **最后更新：** 2026-03-07  
+> **最后更新：** 2026-03-10  
 > **维护者：** 开发团队  
-> **当前状态：** 🚀 **V3.0 "智能审稿增强" 完成（LLM数据核查 + 临床评估 + 稳定性增强）**  
+> **当前状态：** 🚀 **V3.0.1 "性能与体验增强" 完成（4模块并行 + 增量展示 + 导出补全）**  
 > **文档目的：** 快速了解RVW模块状态，为新AI助手提供上下文
 > 
 > **🎉 V3.0 进展（2026-03-07）：**
@@ -14,6 +14,12 @@
 > - ✅ **稳定性增强**：SkillExecutor 使用 Promise.allSettled 实现并行故障隔离
 > - ✅ **部分完成支持**：新增 `partial_completed` 状态 + `errorDetails` 字段，部分模块失败仍展示成功结果
 > - ✅ **前端 4 Tab 报告**：稿约规范性 / 方法学 / 数据验证 / 临床评估，Word 导出全覆盖
+>
+> **⚡ V3.0.1 增强（2026-03-10）：**
+> - ✅ **4模块受控并行**：DataForensics 与 Editorial/Methodology/Clinical 同组并行，`maxConcurrency=4` 控制并发上限
+> - ✅ **增量结果持久化**：每个 Skill 完成即写入任务中间结果，`getTaskDetail` 返回模块级 `reviewProgress`
+> - ✅ **先出先看**：TaskDetail 在审查中即可展示已完成模块（无需等待全流程结束）
+> - ✅ **Word 导出修复**：补齐“数据验证”章节，导出汇总 + 表格明细 + 该表问题列表
 > 
 > **V2.0 进展回顾：**
 > - ✅ L1 算术验证 + L2 统计验证 + L2.5 一致性取证
@@ -34,7 +40,7 @@
 | **商业价值** | ⭐⭐⭐⭐⭐ 极高 |
 | **独立性** | ⭐⭐⭐⭐⭐ 极高（用户群完全不同） |
 | **目标用户** | 期刊初审编辑 |
-| **开发状态** | ✅ **V3.0 完成：4维审查（规范性+方法学+数据验证+临床评估）+ 稳定性增强 + Word导出** |
+| **开发状态** | ✅ **V3.0.1 完成：4维审查并行提速 + 增量结果展示 + Word导出补全** |
 
 ### 核心目标
 
@@ -105,7 +111,7 @@ backend/src/modules/rvw/
 │   ├── clinicalService.ts           # 🆕 V3.0 临床专业评估服务
 │   └── utils.ts                     # 工具函数
 ├── workers/
-│   └── reviewWorker.ts              # pg-boss异步任务处理（V2.0 Skills集成 + V3.0 partial_completed）
+│   └── reviewWorker.ts              # pg-boss异步任务处理（V2.0 Skills集成 + V3.0 partial_completed + V3.0.1 增量落库）
 ├── skills/                          # V2.0 Skills 架构
 │   ├── core/                        # 核心框架（types, registry, executor[allSettled]等）
 │   ├── library/                     # Skill 实现（Forensics[+LLM], Editorial, Methodology, 🆕Clinical）
@@ -125,7 +131,7 @@ frontend-v2/src/modules/rvw/
     ├── Header.tsx                   # 页头（上传按钮）
     ├── Sidebar.tsx                  # 侧边栏导航
     ├── TaskTable.tsx                # 任务列表表格（支持 partial_completed 状态）
-    ├── TaskDetail.tsx               # 任务详情（进度条+报告+Word导出+部分完成警告）
+    ├── TaskDetail.tsx               # 任务详情（进度条+增量展示+Word导出+部分完成警告）
     ├── EditorialReport.tsx          # 稿约规范性报告
     ├── MethodologyReport.tsx        # 方法学评估报告
     ├── ForensicsReport.tsx          # 数据验证报告（含 LLM 核查结果）
@@ -446,6 +452,17 @@ Content-Type: multipart/form-data
 | 前端 partial_completed UI | ✅ 已完成 | 琥珀色警告横幅 + 列表"部分完成"标签 |
 | Word 导出覆盖临床评估 | ✅ 已完成 | 导出报告包含临床专业评估章节 |
 
+### ⚡ V3.0.1 "性能与体验增强" 开发进度（2026-03-10）
+
+| 任务 | 状态 | 说明 |
+|------|------|------|
+| DataForensics 并入并行组 | ✅ 已完成 | 与 Editorial/Methodology/Clinical 同组并行执行 |
+| 并发上限控制 | ✅ 已完成 | Executor 支持按 `maxConcurrency` 分批并行 |
+| 模块完成即持久化 | ✅ 已完成 | Worker 通过 `onSkillComplete` 增量写入 `contextData` |
+| 任务详情增量返回 | ✅ 已完成 | `getTaskDetail` 返回模块结果与 `reviewProgress` |
+| 前端先出先看 | ✅ 已完成 | 审查过程中实时展示已完成 Tab |
+| Word 导出补齐数据验证 | ✅ 已完成 | 导出包含数据验证汇总、表格明细、该表问题列表 |
+
 ### 后续版本（V3.1+）
 
 - [ ] 全面移除评分机制（只列问题，不打分）
@@ -462,7 +479,7 @@ Content-Type: multipart/form-data
 
 ---
 
-**文档版本：** v6.0  
+**文档版本：** v6.1  
-**最后更新：** 2026-03-07  
+**最后更新：** 2026-03-10  
-**当前状态：** 🚀 V3.0 "智能审稿增强" 完成（LLM数据核查 + 临床评估 + 稳定性增强）  
+**当前状态：** 🚀 V3.0.1 "性能与体验增强" 完成（4模块并行 + 增量展示 + 导出补全）  
 **下一步：** V3.1 移除评分机制 + 单模块重试

frontend-v2/src/modules/rvw/components/TaskDetail.tsx

@@ -39,6 +39,7 @@ const getProgressSteps = (selectedAgents: string[]) => {
   const steps = [
     { key: 'upload', label: '上传文档' },
     { key: 'extract', label: '文本提取' },
+    { key: 'forensics', label: '数据验证' },
   ];
   
   if (selectedAgents.includes('editorial')) {
@@ -67,6 +68,17 @@ export default function TaskDetail({ task: initialTask, jobId, onBack }: TaskDet
   const isCompleted = task.status === 'completed' || task.status === 'partial_completed';
   const isPartial = task.status === 'partial_completed';
   const isFailed = task.status === 'failed';
+  const hasIncrementalResult =
+    !!task.editorialReview || !!task.methodologyReview || !!task.forensicsResult || !!task.clinicalReview;
+  const displayReport: ReviewReport | null = report || (hasIncrementalResult
+    ? {
+        ...task,
+        editorialReview: task.editorialReview,
+        methodologyReview: task.methodologyReview,
+        forensicsResult: task.forensicsResult,
+        clinicalReview: task.clinicalReview,
+      }
+    : null);
 
   // 轮询任务状态
   useEffect(() => {
@@ -110,20 +122,20 @@ export default function TaskDetail({ task: initialTask, jobId, onBack }: TaskDet
     }
   }, [isCompleted, task.id, report]);
 
-  // 报告加载后自动设置正确的 Tab
+  // 报告（含增量结果）加载后自动设置可用 Tab
   useEffect(() => {
-    if (report) {
+    if (displayReport) {
-      if (report.editorialReview) {
+      if (displayReport.editorialReview) {
         setActiveTab('editorial');
-      } else if (report.methodologyReview) {
+      } else if (displayReport.methodologyReview) {
         setActiveTab('methodology');
-      } else if (report.forensicsResult) {
+      } else if (displayReport.forensicsResult) {
         setActiveTab('forensics');
-      } else if (report.clinicalReview) {
+      } else if (displayReport.clinicalReview) {
         setActiveTab('clinical');
       }
     }
-  }, [report]);
+  }, [displayReport]);
 
   // 动态获取进度步骤
   const progressSteps = getProgressSteps(task.selectedAgents || ['editorial', 'methodology']);
@@ -132,6 +144,21 @@ export default function TaskDetail({ task: initialTask, jobId, onBack }: TaskDet
   const getStepStatus = (stepKey: string): 'completed' | 'active' | 'pending' => {
     const hasEditorial = task.selectedAgents?.includes('editorial');
     const hasMethodology = task.selectedAgents?.includes('methodology');
+    const hasClinical = task.selectedAgents?.includes('clinical');
+    const skillMap: Record<string, string> = {
+      forensics: 'DataForensicsSkill',
+      editorial: 'EditorialSkill',
+      methodology: 'MethodologySkill',
+      clinical: 'ClinicalAssessmentSkill',
+    };
+    const currentSkill = stepKey in skillMap ? task.reviewProgress?.[skillMap[stepKey]] : undefined;
+
+    if (currentSkill?.status && ['success', 'warning', 'skipped'].includes(currentSkill.status)) {
+      return 'completed';
+    }
+    if (currentSkill?.status && ['error', 'timeout'].includes(currentSkill.status)) {
+      return 'completed';
+    }
     
     if (task.status === 'pending') {
       return stepKey === 'upload' ? 'completed' : 'pending';
@@ -143,7 +170,10 @@ export default function TaskDetail({ task: initialTask, jobId, onBack }: TaskDet
     }
     if (task.status === 'reviewing' || task.status === 'reviewing_editorial') {
       if (['upload', 'extract'].includes(stepKey)) return 'completed';
+      if (stepKey === 'forensics') return 'active';
       if (stepKey === 'editorial' && hasEditorial) return 'active';
+      if (stepKey === 'methodology' && hasMethodology) return 'active';
+      if (stepKey === 'clinical' && hasClinical) return 'active';
       return 'pending';
     }
     if (task.status === 'reviewing_methodology') {
@@ -245,12 +275,20 @@ export default function TaskDetail({ task: initialTask, jobId, onBack }: TaskDet
       );
       
       children.push(new Paragraph({ spacing: { after: 300 } }));
+
+      let sectionIndex = 1;
+      const sectionNumbers = ['一', '二', '三', '四', '五', '六'];
+      const nextSectionTitle = (title: string) => {
+        const current = sectionNumbers[sectionIndex - 1] || `${sectionIndex}`;
+        sectionIndex += 1;
+        return `${current}、${title}`;
+      };
       
       // 稿约规范性评估
       if (report.editorialReview) {
         children.push(
           new Paragraph({
-            text: '一、稿约规范性评估',
+            text: nextSectionTitle('稿约规范性评估'),
             heading: HeadingLevel.HEADING_1,
             spacing: { before: 400, after: 200 },
           })
@@ -313,7 +351,7 @@ export default function TaskDetail({ task: initialTask, jobId, onBack }: TaskDet
       if (report.methodologyReview) {
         children.push(
           new Paragraph({
-            text: '二、方法学评估',
+            text: nextSectionTitle('方法学评估'),
             heading: HeadingLevel.HEADING_1,
             spacing: { before: 400, after: 200 },
           })
@@ -374,13 +412,152 @@ export default function TaskDetail({ task: initialTask, jobId, onBack }: TaskDet
           }
         });
       }
+
+      // 数据验证
+      if (report.forensicsResult) {
+        children.push(
+          new Paragraph({
+            text: nextSectionTitle('数据验证'),
+            heading: HeadingLevel.HEADING_1,
+            spacing: { before: 400, after: 200 },
+          })
+        );
+
+        const summary = report.forensicsResult.summary;
+        children.push(
+          new Paragraph({
+            text: `共识别表格 ${summary.totalTables} 个；发现问题 ${summary.totalIssues} 个（错误 ${summary.errorCount}，警告 ${summary.warningCount}）。`,
+            spacing: { after: 160 },
+          })
+        );
+
+        // 导出表格明细（按 forensicsResult.tables 渲染）
+        if (report.forensicsResult.tables.length > 0) {
+          report.forensicsResult.tables.slice(0, 20).forEach((tableItem, tableIndex) => {
+            children.push(
+              new Paragraph({
+                text: `表${tableIndex + 1}：${tableItem.caption || tableItem.id || '未命名表格'}`,
+                heading: HeadingLevel.HEADING_2,
+                spacing: { before: 180, after: 100 },
+              })
+            );
+
+            const rows: TableRow[] = [];
+            const headers = tableItem.headers || [];
+            const tableData = tableItem.data || [];
+
+            if (headers.length > 0) {
+              rows.push(
+                new TableRow({
+                  children: headers.map((header) => new TableCell({
+                    children: [new Paragraph({ children: [new TextRun({ text: String(header ?? ''), bold: true })] })],
+                  })),
+                })
+              );
+            }
+
+            // 限制导出行数，避免超大表格导致文档膨胀
+            tableData.slice(0, 200).forEach((row) => {
+              rows.push(
+                new TableRow({
+                  children: row.map((cell) => new TableCell({
+                    children: [new Paragraph(String(cell ?? ''))],
+                  })),
+                })
+              );
+            });
+
+            if (rows.length > 0) {
+              children.push(
+                new Table({
+                  width: { size: 100, type: WidthType.PERCENTAGE },
+                  rows,
+                })
+              );
+            } else {
+              children.push(
+                new Paragraph({
+                  text: '（该表未提取到可导出的结构化数据）',
+                  spacing: { after: 80 },
+                })
+              );
+            }
+
+            const tableIssues = (tableItem.issues || []).slice(0, 50);
+            if (tableIssues.length > 0) {
+              children.push(
+                new Paragraph({
+                  children: [new TextRun({ text: '该表问题：', bold: true })],
+                  spacing: { before: 80, after: 40 },
+                })
+              );
+              tableIssues.forEach((issue) => {
+                const level = issue.severity === 'ERROR' ? '错误' : issue.severity === 'WARNING' ? '警告' : '提示';
+                children.push(
+                  new Paragraph({
+                    text: `• [${level}] ${issue.message}`,
+                    indent: { left: 720 },
+                    spacing: { after: 40 },
+                  })
+                );
+              });
+            }
+          });
+
+          if (report.forensicsResult.tables.length > 20) {
+            children.push(
+              new Paragraph({
+                text: `（其余 ${report.forensicsResult.tables.length - 20} 张表格已省略）`,
+                spacing: { before: 80, after: 80 },
+              })
+            );
+          }
+        }
+
+        if (report.forensicsResult.issues.length > 0) {
+          children.push(
+            new Paragraph({
+              children: [new TextRun({ text: '问题清单：', bold: true })],
+              spacing: { after: 80 },
+            })
+          );
+
+          report.forensicsResult.issues.slice(0, 120).forEach((issue) => {
+            const level = issue.severity === 'ERROR' ? '错误' : issue.severity === 'WARNING' ? '警告' : '提示';
+            children.push(
+              new Paragraph({
+                text: `• [${level}] ${issue.message}`,
+                indent: { left: 720 },
+                spacing: { after: 50 },
+              })
+            );
+          });
+
+          if (report.forensicsResult.issues.length > 120) {
+            children.push(
+              new Paragraph({
+                text: `（其余 ${report.forensicsResult.issues.length - 120} 条问题已省略）`,
+                indent: { left: 720 },
+                spacing: { after: 100 },
+              })
+            );
+          }
+        } else {
+          children.push(
+            new Paragraph({
+              children: [new TextRun({ text: '✓ 未发现数据一致性问题', color: '006600' })],
+              indent: { left: 720 },
+              spacing: { after: 100 },
+            })
+          );
+        }
+      }
       
       // 临床专业评估
       if (report.clinicalReview) {
-        const sectionNum = report.methodologyReview ? '三' : '二';
         children.push(
           new Paragraph({
-            text: `${sectionNum}、临床专业评估`,
+            text: nextSectionTitle('临床专业评估'),
             heading: HeadingLevel.HEADING_1,
             spacing: { before: 400, after: 200 },
           })
@@ -552,7 +729,7 @@ export default function TaskDetail({ task: initialTask, jobId, onBack }: TaskDet
               {/* 提示信息 */}
               <div className="text-center text-slate-500 text-sm bg-slate-50 rounded-lg py-4">
                 <p>AI 正在分析您的稿件，这可能需要 1-3 分钟</p>
-                <p className="text-slate-400 mt-1">请耐心等待，完成后将自动显示结果</p>
+                <p className="text-slate-400 mt-1">模块完成后会实时显示，无需等待全部结束</p>
               </div>
             </div>
           )}
@@ -594,8 +771,8 @@ export default function TaskDetail({ task: initialTask, jobId, onBack }: TaskDet
             </div>
           )}
 
-          {/* 完成状态 - 显示报告 */}
+          {/* 增量报告：处理中也可查看已完成模块 */}
-          {isCompleted && report && (
+          {displayReport && (isProcessing || isCompleted || isPartial) && (
             <>
               {/* 审查完成信息 */}
               <div className="bg-gradient-to-br from-indigo-500 to-purple-600 rounded-xl shadow-lg p-5 mb-8 text-white">
@@ -603,10 +780,10 @@ export default function TaskDetail({ task: initialTask, jobId, onBack }: TaskDet
                   <div className="flex items-center gap-3">
                     <CheckCircle className="w-6 h-6 text-white/90" />
                     <div>
-                      <h2 className="text-lg font-semibold">审查完成</h2>
+                      <h2 className="text-lg font-semibold">{isCompleted ? '审查完成' : '已完成模块结果（实时更新）'}</h2>
                       <p className="text-indigo-100 text-sm">
-                        审查用时 {report.durationSeconds ? formatTime(report.durationSeconds) : '-'}
+                        审查用时 {displayReport.durationSeconds ? formatTime(displayReport.durationSeconds) : '-'}
-                        {report.completedAt && ` · ${new Date(report.completedAt).toLocaleString('zh-CN')}`}
+                        {displayReport.completedAt && ` · ${new Date(displayReport.completedAt).toLocaleString('zh-CN')}`}
                       </p>
                     </div>
                   </div>
@@ -615,7 +792,7 @@ export default function TaskDetail({ task: initialTask, jobId, onBack }: TaskDet
 
               {/* Tab切换 */}
               <div className="flex gap-1 bg-slate-200/50 p-1 rounded-lg mb-6 w-fit mx-auto">
-                {report.editorialReview && (
+                {displayReport.editorialReview && (
                   <button 
                     onClick={() => setActiveTab('editorial')}
                     className={`px-6 py-2 rounded-md text-sm transition-all ${
@@ -627,7 +804,7 @@ export default function TaskDetail({ task: initialTask, jobId, onBack }: TaskDet
                     稿约规范性
                   </button>
                 )}
-                {report.methodologyReview && (
+                {displayReport.methodologyReview && (
                   <button 
                     onClick={() => setActiveTab('methodology')}
                     className={`px-6 py-2 rounded-md text-sm transition-all ${
@@ -639,7 +816,7 @@ export default function TaskDetail({ task: initialTask, jobId, onBack }: TaskDet
                     方法学评估
                   </button>
                 )}
-                {report.forensicsResult && (
+                {displayReport.forensicsResult && (
                   <button 
                     onClick={() => setActiveTab('forensics')}
                     className={`px-6 py-2 rounded-md text-sm transition-all ${
@@ -648,10 +825,10 @@ export default function TaskDetail({ task: initialTask, jobId, onBack }: TaskDet
                         : 'font-medium text-slate-500 hover:text-slate-700'
                     }`}
                   >
-                    数据验证 ({report.forensicsResult.summary.totalIssues || 0}个问题)
+                    数据验证 ({displayReport.forensicsResult.summary.totalIssues || 0}个问题)
                   </button>
                 )}
-                {report.clinicalReview && (
+                {displayReport.clinicalReview && (
                   <button 
                     onClick={() => setActiveTab('clinical')}
                     className={`px-6 py-2 rounded-md text-sm transition-all ${
@@ -666,7 +843,7 @@ export default function TaskDetail({ task: initialTask, jobId, onBack }: TaskDet
               </div>
               
               {/* 非 docx 文件无数据验证提示 */}
-              {!report.forensicsResult && (report.editorialReview || report.methodologyReview) && (() => {
+              {!displayReport.forensicsResult && (displayReport.editorialReview || displayReport.methodologyReview) && (() => {
                 const fileName = task.fileName || '';
                 const isPdf = fileName.toLowerCase().endsWith('.pdf');
                 const isDoc = fileName.toLowerCase().endsWith('.doc');
@@ -686,17 +863,17 @@ export default function TaskDetail({ task: initialTask, jobId, onBack }: TaskDet
               })()}
 
               {/* 报告内容 */}
-              {activeTab === 'editorial' && report.editorialReview && (
+              {activeTab === 'editorial' && displayReport.editorialReview && (
-                <EditorialReport data={report.editorialReview} />
+                <EditorialReport data={displayReport.editorialReview} />
               )}
-              {activeTab === 'methodology' && report.methodologyReview && (
+              {activeTab === 'methodology' && displayReport.methodologyReview && (
-                <MethodologyReport data={report.methodologyReview} />
+                <MethodologyReport data={displayReport.methodologyReview} />
               )}
-              {activeTab === 'forensics' && report.forensicsResult && (
+              {activeTab === 'forensics' && displayReport.forensicsResult && (
-                <ForensicsReport data={report.forensicsResult} />
+                <ForensicsReport data={displayReport.forensicsResult} />
               )}
-              {activeTab === 'clinical' && report.clinicalReview && (
+              {activeTab === 'clinical' && displayReport.clinicalReview && (
-                <ClinicalReport data={report.clinicalReview} />
+                <ClinicalReport data={displayReport.clinicalReview} />
               )}
             </>
           )}

frontend-v2/src/modules/rvw/types/index.ts

@@ -42,6 +42,16 @@ export interface ReviewTask {
     timeoutCount: number;
     totalSkills: number;
   };
+  editorialReview?: EditorialReviewResult;
+  methodologyReview?: MethodologyReviewResult;
+  forensicsResult?: ForensicsResult;
+  clinicalReview?: ClinicalReviewResult;
+  reviewProgress?: Record<string, {
+    status: 'success' | 'warning' | 'error' | 'timeout' | 'skipped';
+    executionTime?: number;
+    completedAt?: string;
+    error?: string | null;
+  }>;
   createdAt: string;
   completedAt?: string;
   durationSeconds?: number;