feat: 异步解析大文件WPS报备资料

api/src/files/files-upload-limits.spec.ts

@@ -0,0 +1,18 @@
+import { assertUploadSize } from './files.service';
+
+describe('FilesService report-material upload limits', () => {
+  const workbook = {
+    originalname: '行业报备.xlsx',
+    mimetype: 'application/vnd.openxmlformats-officedocument.spreadsheetml.sheet',
+  };
+
+  it('allows report-material imports up to 100MB without raising the generic file limit', () => {
+    expect(() =>
+      assertUploadSize('report_material_import', { ...workbook, size: 40 * 1024 * 1024 }),
+    ).not.toThrow();
+    expect(() => assertUploadSize('other', { ...workbook, size: 40 * 1024 * 1024 })).toThrow('10MB');
+    expect(() =>
+      assertUploadSize('report_material_import', { ...workbook, size: 101 * 1024 * 1024 }),
+    ).toThrow('100MB');
+  });
+});

api/src/files/files.service.ts

@@ -40,10 +40,12 @@ export class FilesService {
   ) {}
 
   list(tenantId?: string) {
-    return this.prisma.fileObject.findMany({
-      where: tenantId ? { tenantId } : undefined,
-      orderBy: { createdAt: 'desc' },
-    }).then((items) => items.map(serializeFileObject));
+    return this.prisma.fileObject
+      .findMany({
+        where: tenantId ? { tenantId } : undefined,
+        orderBy: { createdAt: 'desc' },
+      })
+      .then((items) => items.map(serializeFileObject));
   }
 
   create(data: CreateFileObjectDto) {
@@ -71,7 +73,7 @@ export class FilesService {
   }
 
   async upload(data: UploadFileDto, file: { originalname: string; mimetype: string; size: number; buffer: Buffer }) {
-    assertUploadSize(file);
+    assertUploadSize(data.purpose, file);
     const fileName = normalizeMultipartFileName(file.originalname);
     const safeName = fileName.replace(/[^\w.\-\u4e00-\u9fa5]/g, '_');
     const objectKey = `${data.prefix ?? data.purpose}/${Date.now()}-${randomUUID()}-${safeName}`;
@@ -87,7 +89,11 @@ export class FilesService {
     });
   }
 
-  async uploadForClient(userId: string | undefined, data: UploadFileDto, file: { originalname: string; mimetype: string; size: number; buffer: Buffer }) {
+  async uploadForClient(
+    userId: string | undefined,
+    data: UploadFileDto,
+    file: { originalname: string; mimetype: string; size: number; buffer: Buffer },
+  ) {
     const tenantId = await this.resolveClientTenantId(userId);
     const prefixRule = CLIENT_UPLOAD_RULES[data.purpose];
     if (!prefixRule || !data.prefix || !prefixRule.test(data.prefix)) {
@@ -135,18 +141,33 @@ export class FilesService {
 
 const IMAGE_UPLOAD_MAX_BYTES = 2 * 1024 * 1024;
 const FILE_UPLOAD_MAX_BYTES = 10 * 1024 * 1024;
+const REPORT_MATERIAL_IMPORT_MAX_BYTES = 100 * 1024 * 1024;
 const IMAGE_FILE_EXTENSION = /\.(?:avif|bmp|gif|heic|heif|jpe?g|png|svg|webp)$/i;
 
-function assertUploadSize(file: { originalname: string; mimetype: string; size: number }) {
+export function assertUploadSize(purpose: string, file: { originalname: string; mimetype: string; size: number }) {
   const image = file.mimetype.toLowerCase().startsWith('image/') || IMAGE_FILE_EXTENSION.test(file.originalname);
-  const limit = image ? IMAGE_UPLOAD_MAX_BYTES : FILE_UPLOAD_MAX_BYTES;
+  const reportMaterialImport = purpose === 'report_material_import';
+  const limit = reportMaterialImport
+    ? REPORT_MATERIAL_IMPORT_MAX_BYTES
+    : image
+      ? IMAGE_UPLOAD_MAX_BYTES
+      : FILE_UPLOAD_MAX_BYTES;
   if (file.size > limit) {
-    throw new BadRequestException(image ? '图片大小不能超过 2MB' : '文件大小不能超过 10MB');
+    throw new BadRequestException(
+      reportMaterialImport
+        ? '报备资料文件大小不能超过 100MB'
+        : image
+          ? '图片大小不能超过 2MB'
+          : '文件大小不能超过 10MB',
+    );
   }
 }
 
 function normalizeMultipartFileName(value: string) {
-  if (![...value].some((character) => character.charCodeAt(0) > 0x7f) || [...value].some((character) => character.charCodeAt(0) > 0xff)) {
+  if (
+    ![...value].some((character) => character.charCodeAt(0) > 0x7f) ||
+    [...value].some((character) => character.charCodeAt(0) > 0xff)
+  ) {
     return value;
   }
 

api/src/report-material-analysis-worker.module.ts

@@ -0,0 +1,13 @@
+import { Module } from '@nestjs/common';
+import { ConfigModule } from '@nestjs/config';
+import { PrismaModule } from './prisma/prisma.module';
+import { ReportMaterialsModule } from './report-materials/report-materials.module';
+
+@Module({
+  imports: [
+    ConfigModule.forRoot({ isGlobal: true, envFilePath: ['.env.local', '.env'] }),
+    PrismaModule,
+    ReportMaterialsModule,
+  ],
+})
+export class ReportMaterialAnalysisWorkerModule {}

api/src/report-material-analysis-worker.ts

@@ -0,0 +1,31 @@
+import 'reflect-metadata';
+import { NestFactory } from '@nestjs/core';
+import { ReportMaterialAnalysisWorkerModule } from './report-material-analysis-worker.module';
+import { ReportMaterialsService } from './report-materials/report-materials.service';
+
+const POLL_INTERVAL_MS = 1000;
+
+async function bootstrap() {
+  if ((process.env.CMPP_PROCESS_ROLE?.trim() || 'report-material-worker') !== 'report-material-worker') {
+    throw new Error('report-material-analysis-worker requires CMPP_PROCESS_ROLE=report-material-worker');
+  }
+  const app = await NestFactory.createApplicationContext(ReportMaterialAnalysisWorkerModule, {
+    logger: ['log', 'warn', 'error'],
+  });
+  app.enableShutdownHooks();
+  const reports = app.get(ReportMaterialsService);
+  await reports.recoverStaleAnalysisJobs();
+  let stopping = false;
+  const stop = () => {
+    stopping = true;
+  };
+  process.once('SIGTERM', stop);
+  process.once('SIGINT', stop);
+  while (!stopping) {
+    const worked = await reports.processNextAnalysisJob();
+    if (!worked) await new Promise((resolve) => setTimeout(resolve, POLL_INTERVAL_MS));
+  }
+  await app.close();
+}
+
+void bootstrap();

api/src/report-materials/import-parser.service.ts

@@ -1,4 +1,4 @@
-import { BadRequestException } from '@nestjs/common';
+import { BadRequestException, NotFoundException } from '@nestjs/common';
 import { Prisma } from '@prisma/client';
 import { extname } from 'node:path';
 import { FilesService } from '../files/files.service';
@@ -13,7 +13,7 @@ import {
   clamp,
   cellText,
 } from './report-materials.helpers';
-import { compatibleImages, loadCompatibleWorkbook } from './workbook-compatibility';
+import { compatibleImagePositions, loadCompatibleWorkbook } from './workbook-compatibility';
 
 /** R4 report-materials domain service composed behind ReportMaterialsService. */
 export class ReportImportParserService {
@@ -63,14 +63,123 @@ export class ReportImportParserService {
     file: { originalname: string; mimetype: string; size: number; buffer: Buffer },
     options: AnalyzeImportOptions,
   ) {
-    if (!options.tenantId) throw new BadRequestException('tenantId is required');
-    if (!options.applicationId) throw new BadRequestException('applicationId is required');
-    await this.smsConfig.getApplication(options.applicationId, options.tenantId);
-    if (!['signature', 'drainage'].includes(options.reportType))
-      throw new BadRequestException('reportType must be signature or drainage');
-    if (extname(file.originalname).toLowerCase() !== '.xlsx' || file.buffer[0] !== 0x50 || file.buffer[1] !== 0x4b)
-      throw new BadRequestException('仅支持有效的 XLSX 文件');
-    const { workbook, wpsImagesBySheet } = await loadCompatibleWorkbook(file.buffer);
+    await this.validateImport(file, options);
+    const analysis = await this.parseAnalysis(file.buffer, options);
+    const sourceFile = await this.files.upload(
+      { tenantId: options.tenantId, purpose: 'report_material_import', prefix: 'report-material-imports' },
+      file,
+    );
+    const batch = await this.createAnalyzedBatch(sourceFile, options, analysis);
+    await this.logAnalyzed(batch.id, sourceFile.fileName, options, analysis.sheetName, analysis.rows.length);
+    return { ...batch, sourceFile, ...analysis };
+  }
+
+  async enqueueImport(
+    file: { originalname: string; mimetype: string; size: number; buffer: Buffer },
+    options: AnalyzeImportOptions,
+  ) {
+    await this.validateImport(file, options);
+    const sourceFile = await this.files.upload(
+      { tenantId: options.tenantId, purpose: 'report_material_import', prefix: 'report-material-imports' },
+      file,
+    );
+    const batch = await this.prisma.reportMaterialImportBatch.create({
+      data: {
+        tenantId: options.tenantId,
+        applicationId: options.applicationId,
+        profileId: options.profileId,
+        fileObjectId: sourceFile.id,
+        fileName: sourceFile.fileName,
+        reportType: options.reportType,
+        status: 'queued',
+        progress: 10,
+        progressStage: '文件已上传，等待解析',
+        sheetName: options.sheetName ?? '',
+        headerRowCount: clamp(options.headerRowCount, 1, 5),
+        dataStartRow: Math.max(options.dataStartRow, clamp(options.headerRowCount, 1, 5) + 1),
+        mapping: [] as Prisma.InputJsonValue,
+        result: { operatorId: options.operatorId } as Prisma.InputJsonValue,
+      },
+    });
+    return this.analysisStatus(batch);
+  }
+
+  async getAnalysisStatus(batchId: string, tenantId: string) {
+    if (!tenantId) throw new BadRequestException('tenantId is required');
+    const batch = await this.prisma.reportMaterialImportBatch.findFirst({ where: { id: batchId, tenantId } });
+    if (!batch) throw new NotFoundException('导入解析任务不存在');
+    return this.analysisStatus(batch);
+  }
+
+  async processQueuedBatch(batchId: string) {
+    const batch = await this.prisma.reportMaterialImportBatch.findUnique({ where: { id: batchId } });
+    if (!batch || batch.status !== 'analyzing') return false;
+    try {
+      await this.progress(batchId, 25, '正在读取工作簿');
+      const { content } = await this.files.getDownload(batch.fileObjectId);
+      const resultData = (batch.result && typeof batch.result === 'object' ? batch.result : {}) as {
+        operatorId?: string;
+      };
+      const options: AnalyzeImportOptions = {
+        tenantId: batch.tenantId,
+        applicationId: batch.applicationId ?? '',
+        reportType: batch.reportType as 'signature' | 'drainage',
+        sheetName: batch.sheetName || undefined,
+        headerRowCount: batch.headerRowCount,
+        dataStartRow: batch.dataStartRow,
+        profileId: batch.profileId ?? undefined,
+        operatorId: resultData.operatorId,
+      };
+      await this.progress(batchId, 30, '正在解析表格与图片位置');
+      const analysis = await this.parseAnalysis(content, options, async (progress, stage) => {
+        await this.progress(batchId, progress, stage);
+      });
+      const updated = await this.prisma.reportMaterialImportBatch.update({
+        where: { id: batchId },
+        data: {
+          status: 'analyzed',
+          progress: 100,
+          progressStage: '解析完成',
+          errorMessage: null,
+          heartbeatAt: new Date(),
+          sheetName: analysis.sheetName,
+          mapping: analysis.suggestedMappings as Prisma.InputJsonValue,
+          preview: {
+            sheets: analysis.sheets,
+            columns: analysis.columns,
+            rows: analysis.rows,
+            imageCount: analysis.imageCount,
+          } as Prisma.InputJsonValue,
+          result: Prisma.JsonNull,
+          rowCount: analysis.rowCount,
+        },
+      });
+      await this.logAnalyzed(batchId, batch.fileName, options, analysis.sheetName, analysis.rows.length);
+      return this.analysisStatus(updated);
+    } catch (error) {
+      const message = error instanceof Error ? error.message : '文件解析失败';
+      await this.prisma.reportMaterialImportBatch.update({
+        where: { id: batchId },
+        data: {
+          status: 'failed',
+          progressStage: '解析失败',
+          errorMessage: message.slice(0, 1000),
+          heartbeatAt: new Date(),
+        },
+      });
+      return false;
+    }
+  }
+
+  private async parseAnalysis(
+    buffer: Buffer,
+    options: AnalyzeImportOptions,
+    onProgress?: (progress: number, stage: string) => void | Promise<void>,
+  ) {
+    const { workbook, wpsImagesBySheet } = await loadCompatibleWorkbook(buffer, {
+      includeImageData: false,
+      onProgress,
+    });
     const profile = options.profileId
       ? await this.prisma.reportMaterialImportProfile.findUnique({
           where: { id: options.profileId },
@@ -82,7 +191,7 @@ export class ReportImportParserService {
     if (!worksheet) throw new BadRequestException('工作簿没有可读取的工作表');
     const headerRowCount = clamp(options.headerRowCount, 1, 5);
     const dataStartRow = Math.max(options.dataStartRow, headerRowCount + 1);
-    const images = compatibleImages(workbook, worksheet, wpsImagesBySheet);
+    const images = compatibleImagePositions(workbook, worksheet, wpsImagesBySheet);
     const columnCount = Math.min(worksheet.columnCount, 200);
     const columns = Array.from({ length: columnCount }, (_, offset) => {
       const sourceColumnIndex = offset + 1;
@@ -110,10 +219,6 @@ export class ReportImportParserService {
       if (Object.values(values).some(Boolean) || imageColumns.length)
         previewRows.push({ rowNumber, values, imageColumns });
     }
-    const sourceFile = await this.files.upload(
-      { tenantId: options.tenantId, purpose: 'report_material_import', prefix: 'report-material-imports' },
-      file,
-    );
     const profileMappings = profile?.columns.map((column) => ({
       sourceHeader: column.sourceHeader,
       sourceHeaderPath: column.sourceHeaderPath ?? undefined,
@@ -128,7 +233,24 @@ export class ReportImportParserService {
     const suggestedMappings = profileMappings?.length
       ? remapProfileColumns(profileMappings, columns)
       : suggestMappings(columns, options.reportType);
-    const batch = await this.prisma.reportMaterialImportBatch.create({
+    await onProgress?.(95, '已生成字段映射与预览');
+    return {
+      sheetName: worksheet.name,
+      sheets: workbook.worksheets.map((sheet) => sheet.name),
+      columns,
+      rows: previewRows,
+      imageCount: images.length,
+      suggestedMappings,
+      rowCount: Math.max(0, worksheet.rowCount - dataStartRow + 1),
+    };
+  }
+
+  private createAnalyzedBatch(
+    sourceFile: { id: string; fileName: string },
+    options: AnalyzeImportOptions,
+    analysis: Awaited<ReturnType<ReportImportParserService['parseAnalysis']>>,
+  ) {
+    return this.prisma.reportMaterialImportBatch.create({
       data: {
         tenantId: options.tenantId,
         applicationId: options.applicationId,
@@ -136,42 +258,91 @@ export class ReportImportParserService {
         fileObjectId: sourceFile.id,
         fileName: sourceFile.fileName,
         reportType: options.reportType,
-        sheetName: worksheet.name,
-        headerRowCount,
-        dataStartRow,
-        mapping: suggestedMappings as Prisma.InputJsonValue,
+        sheetName: analysis.sheetName,
+        headerRowCount: clamp(options.headerRowCount, 1, 5),
+        dataStartRow: Math.max(options.dataStartRow, clamp(options.headerRowCount, 1, 5) + 1),
+        mapping: analysis.suggestedMappings as Prisma.InputJsonValue,
         preview: {
-          sheets: workbook.worksheets.map((sheet) => sheet.name),
-          columns,
-          rows: previewRows,
-          imageCount: images.length,
+          sheets: analysis.sheets,
+          columns: analysis.columns,
+          rows: analysis.rows,
+          imageCount: analysis.imageCount,
         } as Prisma.InputJsonValue,
-        rowCount: Math.max(0, worksheet.rowCount - dataStartRow + 1),
+        rowCount: analysis.rowCount,
       },
     });
+  }
+
+  private async validateImport(
+    file: { originalname: string; size: number; buffer: Buffer },
+    options: AnalyzeImportOptions,
+  ) {
+    if (!options.tenantId) throw new BadRequestException('tenantId is required');
+    if (!options.applicationId) throw new BadRequestException('applicationId is required');
+    await this.smsConfig.getApplication(options.applicationId, options.tenantId);
+    if (!['signature', 'drainage'].includes(options.reportType))
+      throw new BadRequestException('reportType must be signature or drainage');
+    if (extname(file.originalname).toLowerCase() !== '.xlsx' || file.buffer[0] !== 0x50 || file.buffer[1] !== 0x4b)
+      throw new BadRequestException('仅支持有效的 XLSX 文件');
+  }
+
+  private progress(batchId: string, progress: number, progressStage: string) {
+    return this.prisma.reportMaterialImportBatch.update({
+      where: { id: batchId },
+      data: { progress, progressStage, heartbeatAt: new Date() },
+    });
+  }
+
+  private analysisStatus(batch: {
+    id: string;
+    status: string;
+    progress: number;
+    progressStage: string | null;
+    errorMessage: string | null;
+    sheetName: string;
+    preview: Prisma.JsonValue | null;
+    mapping: Prisma.JsonValue;
+  }) {
+    const preview =
+      batch.preview && typeof batch.preview === 'object' && !Array.isArray(batch.preview)
+        ? (batch.preview as Record<string, unknown>)
+        : {};
+    return {
+      id: batch.id,
+      status: batch.status,
+      progress: batch.progress,
+      progressStage: batch.progressStage,
+      errorMessage: batch.errorMessage,
+      sheetName: batch.sheetName || undefined,
+      sheets: preview.sheets ?? [],
+      columns: preview.columns ?? [],
+      rows: preview.rows ?? [],
+      imageCount: preview.imageCount ?? 0,
+      suggestedMappings: batch.mapping ?? [],
+    };
+  }
+
+  private async logAnalyzed(
+    batchId: string,
+    fileName: string,
+    options: AnalyzeImportOptions,
+    sheetName: string,
+    previewCount: number,
+  ) {
     await this.prisma.operationLog.create({
       data: {
         tenantId: options.tenantId,
         userId: options.operatorId,
         action: 'report_material.import_analyzed',
         resource: 'report_material_import',
-        resourceId: batch.id,
+        resourceId: batchId,
         detail: {
-          fileName: sourceFile.fileName,
-          filters: { applicationId: options.applicationId, reportType: options.reportType, sheetName: worksheet.name },
-          successCount: previewRows.length,
+          fileName,
+          filters: { applicationId: options.applicationId, reportType: options.reportType, sheetName },
+          successCount: previewCount,
           failedCount: 0,
         } as Prisma.InputJsonValue,
       },
     });
-    return {
-      ...batch,
-      sourceFile,
-      sheets: workbook.worksheets.map((sheet) => sheet.name),
-      columns,
-      rows: previewRows,
-      imageCount: images.length,
-      suggestedMappings,
-    };
   }
 }

api/src/report-materials/report-materials.contracts.ts

@@ -105,5 +105,6 @@ export type AnalyzeImportOptions = {
 };
 
 export type EmbeddedImage = { row: number; column: number; extension: string; buffer: Buffer };
+export type EmbeddedImageMetadata = { row: number; column: number; extension: string; size: number; target: string };
 
 export type ReportWorkbookFormat = 'excel_drawing' | 'wps_cell_image';

api/src/report-materials/report-materials.controller.ts

@@ -3,6 +3,7 @@ import {
   Body,
   Controller,
   Get,
+  HttpCode,
   Param,
   Post,
   Put,
@@ -93,6 +94,7 @@ export class ReportMaterialsController {
   }
 
   @Post('imports/analyze')
+  @HttpCode(202)
   @UseInterceptors(
     FileInterceptor('file', { limits: { fileSize: 100 * 1024 * 1024, files: 1, fields: 12, parts: 13 } }),
   )
@@ -102,7 +104,7 @@ export class ReportMaterialsController {
     @CurrentSessionUserId() operatorId?: string,
   ) {
     if (!file) throw new BadRequestException('请选择 XLSX 文件');
-    return this.service.analyzeImport(file, {
+    return this.service.enqueueImport(file, {
       tenantId: body.tenantId,
       applicationId: body.applicationId,
       reportType: body.reportType as 'signature' | 'drainage',
@@ -114,6 +116,11 @@ export class ReportMaterialsController {
     });
   }
 
+  @Get('imports/:id/analysis-status')
+  getImportAnalysisStatus(@Param('id') id: string, @Query('tenantId') tenantId: string) {
+    return this.service.getAnalysisStatus(id, tenantId);
+  }
+
   @Put('imports/:id/commit')
   @RequireRecentAuthentication()
   commitImport(@Param('id') id: string, @Body() body: ImportCommitDto, @CurrentSessionUserId() operatorId?: string) {

api/src/report-materials/report-materials.service.spec.ts

@@ -4,6 +4,75 @@ import { ReportMaterialsService } from './report-materials.service';
 import { mappedCorePatchValue, signatureIdentityReportValues } from './report-materials.helpers';
 
 describe('ReportMaterialsService', () => {
+  it('stores a valid workbook as a durable queued analysis job before parsing it', async () => {
+    const prisma = {
+      reportMaterialImportBatch: {
+        create: jest.fn().mockImplementation(({ data }: { data: Record<string, unknown> }) =>
+          Promise.resolve({ id: 'job-1', preview: null, errorMessage: null, ...data }),
+        ),
+      },
+    };
+    const files = { upload: jest.fn().mockResolvedValue({ id: 'file-1', fileName: '行业报备.xlsx' }) };
+    const service = new ReportMaterialsService(
+      prisma as never,
+      files as never,
+      { getApplication: jest.fn().mockResolvedValue({ id: 'app-1' }) } as never,
+    );
+    const result = await service.enqueueImport(
+      { originalname: '行业报备.xlsx', mimetype: 'application/vnd.openxmlformats-officedocument.spreadsheetml.sheet', size: 2, buffer: Buffer.from('PK') },
+      { tenantId: 'tenant-1', applicationId: 'app-1', reportType: 'signature', headerRowCount: 1, dataStartRow: 2 },
+    );
+    expect(files.upload).toHaveBeenCalledWith(
+      expect.objectContaining({ purpose: 'report_material_import', tenantId: 'tenant-1' }),
+      expect.any(Object),
+    );
+    expect(prisma.reportMaterialImportBatch.create).toHaveBeenCalledWith({
+      data: expect.objectContaining({ status: 'queued', progress: 10, applicationId: 'app-1' }),
+    });
+    expect(result).toMatchObject({ id: 'job-1', status: 'queued', progress: 10 });
+  });
+
+  it('only returns analysis progress inside the requested tenant', async () => {
+    const findFirst = jest.fn().mockResolvedValue(null);
+    const service = new ReportMaterialsService(
+      { reportMaterialImportBatch: { findFirst } } as never,
+      {} as never,
+      {} as never,
+    );
+    await expect(service.getAnalysisStatus('job-1', 'tenant-2')).rejects.toThrow('导入解析任务不存在');
+    expect(findFirst).toHaveBeenCalledWith({ where: { id: 'job-1', tenantId: 'tenant-2' } });
+  });
+
+  it('claims only a queued analysis job before handing it to the worker parser', async () => {
+    const prisma = {
+      reportMaterialImportBatch: {
+        findFirst: jest.fn().mockResolvedValue({ id: 'job-1' }),
+        updateMany: jest.fn().mockResolvedValue({ count: 1 }),
+        findUnique: jest.fn().mockResolvedValue(null),
+      },
+    };
+    const service = new ReportMaterialsService(prisma as never, {} as never, {} as never);
+    await expect(service.processNextAnalysisJob()).resolves.toBe(true);
+    expect(prisma.reportMaterialImportBatch.updateMany).toHaveBeenCalledWith({
+      where: { id: 'job-1', status: 'queued' },
+      data: expect.objectContaining({ status: 'analyzing', progress: 15 }),
+    });
+  });
+
+  it('requeues stale analyzing jobs so a restarted worker can resume them', async () => {
+    const updateMany = jest.fn().mockResolvedValue({ count: 2 });
+    const service = new ReportMaterialsService(
+      { reportMaterialImportBatch: { updateMany } } as never,
+      {} as never,
+      {} as never,
+    );
+    await expect(service.recoverStaleAnalysisJobs()).resolves.toEqual({ count: 2 });
+    expect(updateMany).toHaveBeenCalledWith({
+      where: { status: 'analyzing', OR: [{ heartbeatAt: null }, { heartbeatAt: { lt: expect.any(Date) } }] },
+      data: { status: 'queued', progress: 10, progressStage: '等待重新解析', errorMessage: null },
+    });
+  });
+
   it('requires an enterprise application before parsing an import workbook', async () => {
     const service = new ReportMaterialsService({} as never, { upload: jest.fn() } as never, {} as never);
     await expect(

api/src/report-materials/report-materials.service.ts

@@ -34,7 +34,11 @@ export class ReportMaterialsService {
   private readonly channelExport: ReportChannelExportService;
   private readonly batchGeneration: ReportBatchGenerationService;
 
-  constructor(prisma: PrismaService, files: FilesService, smsConfig: SmsConfigService) {
+  constructor(
+    private readonly prisma: PrismaService,
+    files: FilesService,
+    smsConfig: SmsConfigService,
+  ) {
     this.pendingQuery = new ReportPendingQueryService(prisma, files, smsConfig);
     this.officialExport = new ReportOfficialExportService(prisma, files, smsConfig, this.pendingQuery);
     this.importParser = new ReportImportParserService(prisma, files, smsConfig);
@@ -82,6 +86,48 @@ export class ReportMaterialsService {
     return this.importParser.analyzeImport(file, options);
   }
 
+  async enqueueImport(
+    file: { originalname: string; mimetype: string; size: number; buffer: Buffer },
+    options: AnalyzeImportOptions,
+  ) {
+    return this.importParser.enqueueImport(file, options);
+  }
+
+  async getAnalysisStatus(batchId: string, tenantId: string) {
+    return this.importParser.getAnalysisStatus(batchId, tenantId);
+  }
+
+  async recoverStaleAnalysisJobs() {
+    const staleBefore = new Date(Date.now() - 5 * 60 * 1000);
+    return this.prisma.reportMaterialImportBatch.updateMany({
+      where: { status: 'analyzing', OR: [{ heartbeatAt: null }, { heartbeatAt: { lt: staleBefore } }] },
+      data: { status: 'queued', progress: 10, progressStage: '等待重新解析', errorMessage: null },
+    });
+  }
+
+  async processNextAnalysisJob() {
+    const candidate = await this.prisma.reportMaterialImportBatch.findFirst({
+      where: { status: 'queued' },
+      orderBy: { createdAt: 'asc' },
+      select: { id: true },
+    });
+    if (!candidate) return false;
+    const claimed = await this.prisma.reportMaterialImportBatch.updateMany({
+      where: { id: candidate.id, status: 'queued' },
+      data: {
+        status: 'analyzing',
+        progress: 15,
+        progressStage: '后台解析已开始',
+        startedAt: new Date(),
+        heartbeatAt: new Date(),
+        errorMessage: null,
+      },
+    });
+    if (!claimed.count) return true;
+    await this.importParser.processQueuedBatch(candidate.id);
+    return true;
+  }
+
   async commitImport(batchId: string, data: ImportCommitDto) {
     return this.importReview.commitImport(batchId, data);
   }

api/src/report-materials/workbook-compatibility.spec.ts

@@ -1,6 +1,11 @@
 import ExcelJS from 'exceljs';
 import JSZip from 'jszip';
-import { compatibleImages, convertWorkbookOutput, loadCompatibleWorkbook } from './workbook-compatibility';
+import {
+  compatibleImagePositions,
+  compatibleImages,
+  convertWorkbookOutput,
+  loadCompatibleWorkbook,
+} from './workbook-compatibility';
 
 const PNG = Buffer.from(
   'iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAQAAAC1HAwCAAAAC0lEQVR42mP8/x8AAusB9Y9ZrYQAAAAASUVORK5CYII=',
@@ -39,4 +44,15 @@ describe('WPS workbook compatibility', () => {
     const content = Buffer.from(await workbook.xlsx.writeBuffer());
     await expect(loadCompatibleWorkbook(content)).rejects.toThrow('不允许的公式');
   });
+
+  it('reads WPS image positions without inflating image buffers during analysis', async () => {
+    const converted = await convertWorkbookOutput(await standardWorkbook(), 'wps_cell_image');
+    const loaded = await loadCompatibleWorkbook(converted, { includeImageData: false });
+    const wpsImage = loaded.wpsImagesBySheet.get('签名报备')?.[0];
+    expect(wpsImage).toMatchObject({ row: 2, column: 1, extension: 'png', size: PNG.length });
+    expect(wpsImage).not.toHaveProperty('buffer');
+    expect(
+      compatibleImagePositions(loaded.workbook, loaded.workbook.getWorksheet('签名报备')!, loaded.wpsImagesBySheet),
+    ).toEqual([{ row: 2, column: 1 }]);
+  });
 });

api/src/report-materials/workbook-compatibility.ts

@@ -2,7 +2,7 @@ import { BadRequestException } from '@nestjs/common';
 import ExcelJS from 'exceljs';
 import JSZip from 'jszip';
 import { randomUUID } from 'node:crypto';
-import type { EmbeddedImage, ReportWorkbookFormat } from './report-materials.contracts';
+import type { EmbeddedImage, EmbeddedImageMetadata, ReportWorkbookFormat } from './report-materials.contracts';
 import { normalizeImageExtension, readEmbeddedImages } from './report-materials.helpers';
 
 const MAX_WORKBOOK_BYTES = 100 * 1024 * 1024;
@@ -72,9 +72,14 @@ function validateWorkbookValues(workbook: ExcelJS.Workbook, allowedWpsIds: Set<s
   }
 }
 
-async function inspectWpsImages(zip: JSZip) {
+async function inspectWpsImages(zip: JSZip, includeImageData: boolean) {
   const cellImagesXml = await text(zip, 'xl/cellimages.xml');
-  if (!cellImagesXml) return { allowedIds: new Set<string>(), imagesBySheet: new Map<string, EmbeddedImage[]>() };
+  if (!cellImagesXml)
+    return {
+      allowedIds: new Set<string>(),
+      imagesBySheet: new Map<string, Array<EmbeddedImage | EmbeddedImageMetadata>>(),
+      mediaTargets: new Set<string>(),
+    };
   const relXml = await text(zip, 'xl/_rels/cellimages.xml.rels');
   const relTargets = new Map<string, string>();
   for (const match of relXml.matchAll(/<Relationship\b([^>]*)\/?>(?:<\/Relationship>)?/g)) {
@@ -99,13 +104,14 @@ async function inspectWpsImages(zip: JSZip) {
       sheetRelTargets.set(attrs.Id, packagePath(attrs.Target));
   }
   const allowedIds = new Set<string>();
-  const imagesBySheet = new Map<string, EmbeddedImage[]>();
+  const imagesBySheet = new Map<string, Array<EmbeddedImage | EmbeddedImageMetadata>>();
+  let totalImageBytes = 0;
   for (const match of workbookXml.matchAll(/<sheet\b([^>]*)\/?>(?:<\/sheet>)?/g)) {
     const attrs = attributes(match[1]);
     const path = sheetRelTargets.get(attrs['r:id']);
     if (!attrs.name || !path) continue;
     const sheetXml = await text(zip, path);
-    const images: EmbeddedImage[] = [];
+    const images: Array<EmbeddedImage | EmbeddedImageMetadata> = [];
     for (const cell of sheetXml.matchAll(/<c\b([^>]*)>([\s\S]*?)<\/c>/g)) {
       const formulaText = decodeXml(cell[2].match(/<f(?:\s[^>]*)?>([\s\S]*?)<\/f>/)?.[1]?.trim() ?? '');
       if (!formulaText) continue;
@@ -118,20 +124,31 @@ async function inspectWpsImages(zip: JSZip) {
       if (!imageEntry) throw new BadRequestException('WPS单元格图片文件缺失');
       const extension = normalizeImageExtension(target.split('.').pop() ?? 'png');
       if (!['png', 'jpeg', 'gif'].includes(extension)) throw new BadRequestException('WPS单元格图片格式不受支持');
-      const imageBuffer = await imageEntry.async('nodebuffer');
-      if (imageBuffer.length > MAX_IMAGE_BYTES) throw new BadRequestException('单张WPS图片不能超过20MB');
-      if (!validImageSignature(extension, imageBuffer)) throw new BadRequestException('WPS单元格图片内容与格式不匹配');
-      images.push({ ...cellPosition, extension, buffer: imageBuffer });
+      const size = Number(
+        (imageEntry as unknown as { _data?: { uncompressedSize?: number } })._data?.uncompressedSize ?? 0,
+      );
+      if (size > MAX_IMAGE_BYTES) throw new BadRequestException('单张WPS图片不能超过20MB');
+      totalImageBytes += size;
+      if (includeImageData) {
+        const imageBuffer = await imageEntry.async('nodebuffer');
+        if (!validImageSignature(extension, imageBuffer))
+          throw new BadRequestException('WPS单元格图片内容与格式不匹配');
+        images.push({ ...cellPosition, extension, buffer: imageBuffer });
+      } else {
+        images.push({ ...cellPosition, extension, size, target });
+      }
       allowedIds.add(formula[1]);
     }
     imagesBySheet.set(attrs.name, images);
   }
-  const totalImageBytes = [...imagesBySheet.values()].flat().reduce((sum, image) => sum + image.buffer.length, 0);
   if (totalImageBytes > MAX_TOTAL_IMAGE_BYTES) throw new BadRequestException('WPS图片总量不能超过300MB');
-  return { allowedIds, imagesBySheet };
+  return { allowedIds, imagesBySheet, mediaTargets: new Set(imageTargets.values()) };
 }
 
-export async function loadCompatibleWorkbook(buffer: Buffer) {
+export async function loadCompatibleWorkbook(
+  buffer: Buffer,
+  options: { includeImageData?: boolean; onProgress?: (progress: number, stage: string) => void | Promise<void> } = {},
+) {
   if (buffer.length > MAX_WORKBOOK_BYTES) throw new BadRequestException('导入文件不能超过100MB');
   let zip: JSZip;
   try {
@@ -139,6 +156,7 @@ export async function loadCompatibleWorkbook(buffer: Buffer) {
   } catch {
     throw new BadRequestException('仅支持有效的 XLSX 文件');
   }
+  await options.onProgress?.(35, '已读取工作簿压缩包');
   const expandedBytes = Object.values(zip.files).reduce(
     (sum, entry) =>
       sum + Number((entry as unknown as { _data?: { uncompressedSize?: number } })._data?.uncompressedSize ?? 0),
@@ -148,17 +166,32 @@ export async function loadCompatibleWorkbook(buffer: Buffer) {
   if (expandedBytes > MAX_EXPANDED_BYTES) throw new BadRequestException('工作簿解压后超过500MB安全限制');
   if (Object.keys(zip.files).some((name) => /(^|\/)(vbaProject|externalLinks|embeddings|activeX)(\/|\.)/i.test(name)))
     throw new BadRequestException('工作簿包含不允许的外部对象或宏');
-  const wps = await inspectWpsImages(zip);
+  const wps = await inspectWpsImages(zip, options.includeImageData !== false);
+  await options.onProgress?.(55, '已定位WPS单元格图片');
+  let excelJsInput = buffer;
+  if (options.includeImageData === false && wps.mediaTargets.size) {
+    for (const target of wps.mediaTargets) zip.remove(target);
+    zip.remove('xl/cellimages.xml');
+    zip.remove('xl/_rels/cellimages.xml.rels');
+    excelJsInput = await zip.generateAsync({
+      type: 'nodebuffer',
+      compression: 'DEFLATE',
+      compressionOptions: { level: 1 },
+    });
+    await options.onProgress?.(70, '已剥离解析阶段无需加载的WPS图片数据');
+  }
   const workbook = new ExcelJS.Workbook();
-  await workbook.xlsx.load(buffer as never);
+  await workbook.xlsx.load(excelJsInput as never);
+  await options.onProgress?.(80, '已解析工作表内容');
   validateWorkbookValues(workbook, wps.allowedIds);
+  await options.onProgress?.(90, '已完成工作簿安全校验');
   return { workbook, wpsImagesBySheet: wps.imagesBySheet };
 }
 
 export function compatibleImages(
   workbook: ExcelJS.Workbook,
   worksheet: ExcelJS.Worksheet,
-  wpsImagesBySheet: Map<string, EmbeddedImage[]>,
+  wpsImagesBySheet: Map<string, Array<EmbeddedImage | EmbeddedImageMetadata>>,
 ) {
   const merged = new Map<string, EmbeddedImage>();
   const wpsImages = wpsImagesBySheet.get(worksheet.name) ?? [];
@@ -169,10 +202,29 @@ export function compatibleImages(
     if (!validImageSignature(extension, image.buffer)) throw new BadRequestException('工作簿图片内容与格式不匹配');
     merged.set(`${image.row}:${image.column}`, image);
   }
-  for (const image of wpsImages) merged.set(`${image.row}:${image.column}`, image);
+  for (const image of wpsImages) {
+    if (!('buffer' in image)) throw new BadRequestException('图片元数据不能用于提交导入');
+    merged.set(`${image.row}:${image.column}`, image);
+  }
   return [...merged.values()];
 }
 
+export function compatibleImagePositions(
+  workbook: ExcelJS.Workbook,
+  worksheet: ExcelJS.Worksheet,
+  wpsImagesBySheet: Map<string, Array<EmbeddedImage | EmbeddedImageMetadata>>,
+) {
+  const wpsImages = wpsImagesBySheet.get(worksheet.name) ?? [];
+  const positions = new Map<string, { row: number; column: number }>();
+  for (const image of readEmbeddedImages(workbook, worksheet)) {
+    if (wpsImages.length && image.buffer.length <= 128) continue;
+    positions.set(`${image.row}:${image.column}`, { row: image.row, column: image.column });
+  }
+  for (const image of wpsImages)
+    positions.set(`${image.row}:${image.column}`, { row: image.row, column: image.column });
+  return [...positions.values()];
+}
+
 export async function convertWorkbookOutput(content: Buffer, outputFormat: ReportWorkbookFormat = 'excel_drawing') {
   if (outputFormat === 'excel_drawing') return content;
   if (outputFormat !== 'wps_cell_image') throw new BadRequestException('不支持的报备文件格式');