feat(ee): PDF import (#2142)

* feat: replace pdfjs-dist with firecrawl-pdf-inspector

* use modified firecrawl-pdf-inspector

* feat: pdf import

* increase single file upload size limit

* use npm package

* sync

* update package

apps/server/src/common/features.ts

@@ -8,6 +8,7 @@ export const Feature = {
   AI: 'ai',
   CONFLUENCE_IMPORT: 'import:confluence',
   DOCX_IMPORT: 'import:docx',
+  PDF_IMPORT: 'import:pdf',
   ATTACHMENT_INDEXING: 'attachment:indexing',
   SECURITY_SETTINGS: 'security:settings',
   MCP: 'mcp',

apps/server/src/integrations/import/import.controller.ts

@@ -51,9 +51,9 @@ export class ImportController {
     @AuthUser() user: User,
     @AuthWorkspace() workspace: Workspace,
   ) {
-    const validFileExtensions = ['.md', '.html', '.docx'];
+    const validFileExtensions = ['.md', '.html', '.docx', '.pdf'];
 
-    const maxFileSize = bytes('20mb');
+    const maxFileSize = bytes('30mb');
 
     let file = null;
     try {
@@ -102,6 +102,7 @@ export class ImportController {
       '.md': 'markdown',
       '.html': 'html',
       '.docx': 'docx',
+      '.pdf': 'pdf',
     };
 
     if (createdPage) {

apps/server/src/integrations/import/services/import.service.ts

@@ -63,7 +63,10 @@ export class ImportService {
     let createdPage = null;
 
     // For DOCX, we need the page ID upfront so images can reference it
-    const pageId = fileExtension === '.docx' ? uuid7() : undefined;
+    const pageId =
+      fileExtension === '.docx' || fileExtension === '.pdf'
+        ? uuid7()
+        : undefined;
 
     try {
       if (fileExtension.endsWith('.md')) {
@@ -78,6 +81,14 @@ export class ImportService {
           pageId,
           userId,
         );
+      } else if (fileExtension.endsWith('.pdf')) {
+        prosemirrorState = await this.processPdf(
+          fileBuffer,
+          workspaceId,
+          spaceId,
+          pageId,
+          userId,
+        );
       }
     } catch (err) {
       const message = 'Error processing file content';
@@ -156,7 +167,7 @@ export class ImportService {
     let DocxImportModule: any;
     try {
       // eslint-disable-next-line @typescript-eslint/no-require-imports
-      DocxImportModule = require('./../../../ee/docx-import/docx-import.service');
+      DocxImportModule = require('./../../../ee/document-import/docx-import.service');
     } catch (err) {
       this.logger.error(
         'DOCX import requested but EE module not bundled in this build',
@@ -182,6 +193,42 @@ export class ImportService {
     return this.processHTML(html);
   }
 
+  async processPdf(
+    fileBuffer: Buffer,
+    workspaceId: string,
+    spaceId: string,
+    pageId: string,
+    userId: string,
+  ): Promise<any> {
+    let PdfImportModule: any;
+    try {
+      // eslint-disable-next-line @typescript-eslint/no-require-imports
+      PdfImportModule = require('./../../../ee/document-import/pdf-import.service');
+    } catch (err) {
+      this.logger.error(
+        'PDF import requested but EE module not bundled in this build',
+      );
+      throw new BadRequestException(
+        'This feature requires a valid enterprise license.',
+      );
+    }
+
+    const pdfImportService = this.moduleRef.get(
+      PdfImportModule.PdfImportService,
+      { strict: false },
+    );
+
+    const html = await pdfImportService.convertPdfToHtml(
+      fileBuffer,
+      workspaceId,
+      spaceId,
+      pageId,
+      userId,
+    );
+
+    return this.processHTML(html);
+  }
+
   async createYdoc(prosemirrorJson: any): Promise<Buffer | null> {
     if (prosemirrorJson) {
       // this.logger.debug(`Converting prosemirror json state to ydoc`);