iistwin/server/documents/docx-importer.service.ts

import mammoth from "mammoth";
import fs from "fs/promises";
import path from "path";
import { uploadToS3, uploadToS3FromFile } from "../utils/s3";
import * as storage from "../storage/documents.storage";
import { db } from "../db";
import { fileUploads } from "@shared/schema";
import { DocumentTemplateService } from "./template.service";
import type { MulterFile } from "./asset.service";
import { convertEmfToPng } from "./external-pdf.service";

export interface ImportDocxOptions {
  name?: string;
  description?: string;
  categoryId?: number;
  formId?: number;
}

export class DocxImporterService {
  constructor(private templateService: DocumentTemplateService) {}

  async importDocx(
    file: MulterFile,
    organizationId: number,
    userId: number,
    options: ImportDocxOptions = {},
  ) {
    if (!file.path) {
      throw new Error("Файл не сохранён на диск");
    }

    const templateName = options.name || this.sanitizeName(file.originalname);

    const { html, sourceDocxFileKey, warnings } = await this.processDocxFile(
      file,
      templateName,
      organizationId,
      userId,
    );

    const template = await this.templateService.create({
      organizationId,
      name: templateName,
      description: options.description,
      categoryId: options.categoryId,
      formId: options.formId,
      content: html,
      status: "draft",
      pageOrientation: "portrait",
      pageSize: "A4",
      margins: { top: 1440, left: 1440, right: 1440, bottom: 1440 },
      assetPlacements: [],
      sourceDocxFileKey,
      createdBy: userId,
    });

    const variables = await this.templateService.autoCreateVariablesFromContent(
      template.id,
      organizationId,
    );

    return {
      template: { ...template, variables },
      variables,
      warnings,
    };
  }

  async replaceDocx(
    file: MulterFile,
    templateId: number,
    organizationId: number,
    userId: number,
  ) {
    if (!file.path) {
      throw new Error("Файл не сохранён на диск");
    }

    const template = await this.templateService.getByIdRaw(templateId, organizationId);
    if (!template) throw new Error("Шаблон не найден");

    const templateName = this.sanitizeName(file.originalname);

    const { html, sourceDocxFileKey, warnings } = await this.processDocxFile(
      file,
      templateName,
      organizationId,
      userId,
    );

    await this.templateService.update(templateId, organizationId, {
      name: template.name,
      content: html,
      sourceDocxFileKey,
    });

    const variables = await this.templateService.autoCreateVariablesFromContent(
      templateId,
      organizationId,
    );

    const updated = await this.templateService.getById(templateId, organizationId);

    return {
      template: updated!,
      variables,
      warnings,
    };
  }

  private async processDocxFile(
    file: MulterFile,
    templateName: string,
    organizationId: number,
    userId: number,
  ): Promise<{ html: string; sourceDocxFileKey: string; warnings: string[] }> {
    const result = await mammoth.convertToHtml(
      { path: file.path },
      {
        convertImage: mammoth.images.imgElement(async (image) => {
          const buffer = await image.read("base64");
          let binary = Buffer.from(buffer, "base64");
          let contentType = image.contentType || "image/png";
          let ext = this.extFromContentType(contentType);

          // Конвертируем EMF/WMF в PNG через внешний document-worker
          if (contentType === "image/x-emf" || contentType === "image/x-wmf") {
            const png = await convertEmfToPng(binary);
            if (!png) {
              throw new Error("Сервис конвертации изображений временно недоступен. Не удалось конвертировать EMF/WMF в PNG.");
            }
            binary = png;
            contentType = "image/png";
            ext = "png";
          }

          const originalName = `image-${Date.now()}-${Math.random().toString(36).slice(2)}.${ext}`;

          const { key } = await uploadToS3(binary, originalName, contentType);

          await storage.createOrganizationAsset({
            organizationId,
            type: "image",
            label: originalName,
            fileKey: key,
            userId,
          });

          await db.insert(fileUploads).values({
            organizationId,
            uploadedBy: userId,
            fileKey: key,
            originalName,
            sizeBytes: binary.length,
            taskId: null,
            fieldId: null,
          }).onConflictDoNothing();

          return { src: `/api/files/${key}` };
        }),
      },
    );

    const html = this.sanitizeHtml(result.value);

    // Сохраняем исходный DOCX как шаблон
    const { key: sourceDocxFileKey } = await uploadToS3FromFile(
      file.path,
      `${templateName}.docx`,
      "application/vnd.openxmlformats-officedocument.wordprocessingml.document",
    );

    await fs.unlink(file.path).catch(() => {});

    return {
      html,
      sourceDocxFileKey,
      warnings: result.messages.map((m) => m.message),
    };
  }

  private sanitizeName(originalname: string): string {
    const base = path.basename(originalname, path.extname(originalname));
    return base.replace(/[\/\\?%*:|<>"]/g, "-").trim() || "Импортированный шаблон";
  }

  private extFromContentType(contentType: string | undefined): string {
    switch (contentType) {
      case "image/png":
        return "png";
      case "image/jpeg":
        return "jpg";
      case "image/gif":
        return "gif";
      case "image/webp":
        return "webp";
      case "image/svg+xml":
        return "svg";
      default:
        return "png";
    }
  }

  private sanitizeHtml(html: string): string {
    // mammoth может генерировать <img> без закрывающего слеша и с \n в src.
    // Приводим к виду, удобному для TipTap.
    return html
      .replace(/\s+/g, " ")
      .replace(/<img\s+([^>]*?)>/gi, (match, attrs) => {
        const srcMatch = /src="([^"]+)"/.exec(attrs);
        const src = srcMatch ? srcMatch[1] : "";
        const altMatch = /alt="([^"]*)"/.exec(attrs);
        const alt = altMatch ? altMatch[1] : "";
        return `<img src="${src}" alt="${alt}">`;
      })
      .trim();
  }
}