Add ppt deps

server/src/services/documentParser.ts

@@ -1,4 +1,8 @@
 import officeParser from 'officeparser';
+import WordExtractor from 'word-extractor';
+import * as XLSX from 'xlsx';
+import * as PPT from 'ppt';
+import * as CFB from 'cfb';
 
 export const GEMINI_NATIVE_TYPES = [
   'application/pdf',
@@ -30,20 +34,22 @@ export const TYPES_TO_CONVERT_TO_PLAIN_TEXT = [
   'application/x-yaml',
 ];
 
-// MIME types that officeparser can extract text from
+// MIME types that officeparser can extract text from (modern Office formats)
 export const OFFICEPARSER_TYPES = [
   'application/vnd.openxmlformats-officedocument.wordprocessingml.document',   // .docx
   'application/vnd.openxmlformats-officedocument.presentationml.presentation', // .pptx
   'application/vnd.openxmlformats-officedocument.spreadsheetml.sheet',         // .xlsx
-  'application/msword',                                                         // .doc
-  'application/vnd.ms-powerpoint',                                              // .ppt
-  'application/vnd.ms-excel',                                                   // .xls
   'application/vnd.oasis.opendocument.text',                                   // .odt
   'application/vnd.oasis.opendocument.presentation',                           // .odp
   'application/vnd.oasis.opendocument.spreadsheet',                            // .ods
   'application/rtf',                                                            // .rtf
 ];
 
+// Legacy Office formats requiring specialized parsers
+export const LEGACY_DOC_TYPE = 'application/msword';                            // .doc (word-extractor)
+export const LEGACY_XLS_TYPE = 'application/vnd.ms-excel';                      // .xls (xlsx/SheetJS)
+export const LEGACY_PPT_TYPE = 'application/vnd.ms-powerpoint';                 // .ppt (ppt/SheetJS)
+
 // Image types that can use OCR for text extraction
 export const OCR_CAPABLE_TYPES = [
   'image/jpeg',
@@ -52,7 +58,14 @@ export const OCR_CAPABLE_TYPES = [
   'image/webp'
 ];
 
-export const SUPPORTED_TYPES = [...GEMINI_NATIVE_TYPES, ...OFFICEPARSER_TYPES, ...TYPES_TO_CONVERT_TO_PLAIN_TEXT];
+export const SUPPORTED_TYPES = [
+  ...GEMINI_NATIVE_TYPES, 
+  ...OFFICEPARSER_TYPES, 
+  ...TYPES_TO_CONVERT_TO_PLAIN_TEXT,
+  LEGACY_DOC_TYPE,
+  LEGACY_XLS_TYPE,
+  LEGACY_PPT_TYPE,
+];
 
 const EXTENSION_TO_MIME: Record<string, string> = {
   '.txt': 'text/plain',
@@ -129,6 +142,18 @@ export function needsPlainTextConversion(mimeType: string): boolean {
   return TYPES_TO_CONVERT_TO_PLAIN_TEXT.includes(mimeType);
 }
 
+export function needsWordExtractor(mimeType: string): boolean {
+  return mimeType === LEGACY_DOC_TYPE;
+}
+
+export function needsSheetJS(mimeType: string): boolean {
+  return mimeType === LEGACY_XLS_TYPE;
+}
+
+export function needsSheetJSPPT(mimeType: string): boolean {
+  return mimeType === LEGACY_PPT_TYPE;
+}
+
 async function extractWithOfficeParser(buffer: Buffer, useOcr: boolean = false): Promise<string> {
   const config = useOcr ? {
     extractAttachments: true,
@@ -150,6 +175,34 @@ async function extractWithOfficeParser(buffer: Buffer, useOcr: boolean = false):
   return text;
 }
 
+async function extractWithWordExtractor(buffer: Buffer): Promise<string> {
+  const extractor = new WordExtractor();
+  const doc = await extractor.extract(buffer);
+  return doc.getBody();
+}
+
+function extractWithSheetJS(buffer: Buffer): string {
+  const workbook = XLSX.read(buffer, { type: 'buffer' });
+  const textParts: string[] = [];
+  
+  for (const sheetName of workbook.SheetNames) {
+    const sheet = workbook.Sheets[sheetName];
+    const csv = XLSX.utils.sheet_to_csv(sheet);
+    if (csv.trim()) {
+      textParts.push(`--- ${sheetName} ---\n${csv}`);
+    }
+  }
+  
+  return textParts.join('\n\n');
+}
+
+function extractWithSheetJSPPT(buffer: Buffer): string {
+  const cfb = CFB.read(buffer, { type: 'buffer' });
+  const pres = PPT.parse_pptcfb(cfb);
+  const textArray = PPT.utils.to_text(pres);
+  return textArray.join('\n\n');
+}
+
 export async function processDocument(
   buffer: Buffer,
   mimeType: string,
@@ -188,6 +241,36 @@ export async function processDocument(
     };
   }
 
+  // Legacy .doc files (Word 97-2003)
+  if (needsWordExtractor(mimeType)) {
+    const text = await extractWithWordExtractor(buffer);
+    return {
+      type: 'text',
+      content: text,
+      mimeType: 'text/plain'
+    };
+  }
+
+  // Legacy .xls files (Excel 97-2003)
+  if (needsSheetJS(mimeType)) {
+    const text = extractWithSheetJS(buffer);
+    return {
+      type: 'text',
+      content: text,
+      mimeType: 'text/plain'
+    };
+  }
+
+  // Legacy .ppt files (PowerPoint 97-2003)
+  if (needsSheetJSPPT(mimeType)) {
+    const text = extractWithSheetJSPPT(buffer);
+    return {
+      type: 'text',
+      content: text,
+      mimeType: 'text/plain'
+    };
+  }
+
   if (needsPlainTextConversion(mimeType)) {
     return {
       type: 'text',

server/src/types/legacy-office.d.ts

@@ -0,0 +1,46 @@
+// Type declarations for legacy Office format parsers
+
+declare module 'ppt' {
+  interface PPTPresentation {
+    docs: Array<{
+      slideList: string[];
+    }>;
+    slides: Array<{
+      drawing: {
+        groupShape: Array<{
+          clientTextbox?: {
+            t: string;
+          };
+        }>;
+      };
+    }>;
+  }
+
+  interface PPTUtils {
+    to_text(pres: PPTPresentation): string[];
+  }
+
+  export function readFile(filename: string, opts?: { WTF?: number; dump?: number }): PPTPresentation;
+  export function parse_pptcfb(cfb: CFB.Container, opts?: object): PPTPresentation;
+  export const utils: PPTUtils;
+}
+
+declare module 'cfb' {
+  interface CFBEntry {
+    name: string;
+    type: number;
+    content?: Buffer;
+  }
+
+  interface Container {
+    FileIndex: CFBEntry[];
+    FullPaths: string[];
+  }
+
+  interface ReadOptions {
+    type?: 'file' | 'buffer' | 'base64' | 'binary' | 'array';
+  }
+
+  export function read(data: Buffer | string | ArrayBuffer, opts?: ReadOptions): Container;
+  export function parse(data: Buffer | number[]): Container;
+}