Dcument extraction documented

Author: oksbwn

docs/feature-reference.md

@@ -78,9 +78,11 @@ During transfer:
 ### Diagrams & Media Gallery
 
 Open via **Workspace -> Diagrams & Media Gallery** (`Ctrl/Cmd + Alt + M`) or Command Palette:
-- Catalogs all used diagrams (inline Mermaid, Draw.io, Excalidraw, Wireframe UI prototypes), images, videos, audio, and PDF documents in the workspace.
+- Catalogs all used diagrams (inline Mermaid, Draw.io, Excalidraw, Wireframe UI prototypes), images, videos, audio, and PDF/Office documents in the workspace.
 - **Physical Disk Scanner & Orphan Detection**: Scans workspace folders (`media/`, `assets/`, `images/`, `media/wireframes/`, `media/draw.io/`) to identify files present on disk that have zero note references.
 - **Unused Media Filter**: Quickly isolate unreferenced assets with the `⚠️ Unused / Orphans` filter tab.
+- **Automated Document Extraction**: Extracts and caches clean structured Markdown from Word (`.docx`), Excel (`.xlsx`, `.xls`, `.csv`), PowerPoint (`.pptx`), and PDF (`.pdf`) files with table preservation and spatial clustering.
+- **Tabbed Media & Extracted Viewer**: Switch between native preview and formatted extracted reading view with in-document search, line-numbered raw view, copy/download tools, and metadata chips.
 - **Audio & Media Preview**: Embedded audio players, video thumbnails, and transcript drawer with key points and action items.
 - **UI Prototypes & Wireframe Studio**: View, inspect references, and launch Wireframe Studio to edit interactive UI mockups directly.
 - Reuses Knowledge Graph layout aesthetics with collapsible category filters, search, and stat indicators.

docs/workspace/media.md

@@ -39,10 +39,30 @@ Keep your assets tidy using the Media Health Dashboard:
 ## 4. Diagrams & Media Asset Manager
 
 Notely provides a dedicated full-screen asset manager under **Workspace > Diagrams & Media** (shortcut: `Ctrl+Alt+M` or `Cmd+Alt+M`):
-- **Dense Asset Table Layout**: High-density table displaying Asset Name, Type Badge, File Size, Reference Status, and Quick Action buttons.
+- **Dense Asset Table Layout**: High-density table displaying Asset Name, Type Badge, File Size, Reference Status, Extraction Status, and Quick Action buttons.
 - **Fixed Sidebar & Category Filters**: Left sidebar (260px) with real-time counters for Diagrams, UI Prototypes, Images, PDFs, Videos, Audio, Transcripts, and Documents.
 - **Physical Disk Scanner & Orphan Detection**: Discovers assets on disk across `media/`, `assets/`, and `images/`. Unreferenced assets are highlighted with the `⚠️ Unused / Orphans` badge and filter tab.
 - **Asset Bundles**: Audio and video recordings are organized into bundle folders (`media/audio/<id>/` and `media/video/<id>/`) containing the primary recording alongside companion `transcript.json`.
 - **Integrated Full-Screen Media Viewer**: Standardized modal viewer with 100% full-width layout, subtitle/transcript side drawer, playback speed controls, and copy/download tools.
 - **Safe Asset Deletion**: Remove unwanted assets with safety confirmation modals that check note references and warn before removing actively linked files.
 - **Referenced Notes & Line Jumping**: Inspect reference pills to view note names, line numbers, and markdown context snippets, with 1-click navigation to jump directly to the editor position.
+
+---
+
+## 5. Document Text Extraction & Caching
+
+Notely features an automated, background document extraction and indexing engine for office documents and PDFs:
+
+### Supported Formats
+- **Word Documents (`.docx`)**: OpenXML parsing extracting headings, paragraphs, bulleted/numbered lists, and tables as standard Markdown.
+- **Spreadsheets (`.xlsx`, `.xls`, `.csv`)**: SheetJS-powered table extraction preserving multi-sheet workbooks (`### Sheet: Name`) into GFM Markdown tables.
+- **Presentations (`.pptx`)**: Slide-by-slide text hierarchy extraction (`### Slide N: Title`) including presenter speaker notes.
+- **PDF Documents (`.pdf`)**: Spatial line clustering and page-segmented Markdown extraction (`### Page N`).
+
+### Unified Tabbed Preview
+When inspecting or previewing supported documents in the gallery or editor:
+- **`[ Preview | Extracted Text ]` Switcher**: Seamlessly toggle between native rendering and the extracted text view.
+- **Formatted Reading View**: Formatted paper sheet view with typography, scrollable tables, reading time estimation, and metadata pills (page count, word count, SHA-256 content hash).
+- **Markdown Source View**: Synchronized line-numbered raw markdown editor with copy, download, and re-extraction actions.
+- **In-Document Search**: Instant search bar with occurrence counting and keyword highlights across both reading view and raw source lines.
+- **Persistent Caching**: Document hashes and extracted text are stored in `.notes-app/documents.db` and `.notes-app/extracted/` so subsequent reads are instant without re-parsing.

package.json

@@ -50,6 +50,7 @@
     "node-pty": "^1.1.0",
     "onnxruntime-node": "^1.27.0",
     "onnxruntime-web": "^1.27.0",
+    "pdfjs-dist": "^4.3.136",
     "simple-git": "^3.36.0",
     "tree-sitter-wasms": "^0.1.13",
     "uuid": "^14.0.1",
@@ -88,7 +89,6 @@
     "markdownlint-cli2": "^0.18.1",
     "mermaid": "^11.6.0",
     "nspell": "^2.1.5",
-    "pdfjs-dist": "^4.3.136",
     "png-to-ico": "^2.1.8",
     "pngjs": "^7.0.0",
     "prettier": "^3.4.0",

tests/documentExtractors.test.js

@@ -8,6 +8,8 @@ import * as XLSX from 'xlsx';
 const { extractDocx } = require('../electron/lib/extractors/DocxExtractor.cjs');
 const { extractPptx } = require('../electron/lib/extractors/PptxExtractor.cjs');
 const { extractSpreadsheet } = require('../electron/lib/extractors/SpreadsheetExtractor.cjs');
+const { extractPdf } = require('../electron/lib/extractors/PdfExtractor.cjs');
+const { extractDocument, isSupportedDocument, getDocumentMimeType } = require('../electron/lib/extractors/DocumentExtractor.cjs');
 const { DocumentCacheStore } = require('../electron/lib/documents/DocumentCacheStore.cjs');
 const { DocumentExtractionService } = require('../electron/services/DocumentExtractionService.cjs');
 
@@ -224,4 +226,51 @@ describe('Document Extractors & Cache Pipeline', () => {
       service.close();
     });
   });
+
+  describe('PdfExtractor & Error Handling', () => {
+    it('handles non-existent PDF files gracefully', async () => {
+      const result = await extractPdf(path.join(tempDir, 'missing.pdf'));
+      expect(result.success).toBe(false);
+      expect(result.status).toBe('failed');
+      expect(result.error).toContain('File not found');
+    });
+
+    it('handles corrupted/empty PDF files gracefully', async () => {
+      const corruptedPath = path.join(tempDir, 'corrupt.pdf');
+      fs.writeFileSync(corruptedPath, 'Not a real PDF stream header');
+      const result = await extractPdf(corruptedPath);
+      expect(result.success).toBe(false);
+      expect(result.status).toBe('failed');
+    });
+  });
+
+  describe('DocumentExtractor Dispatcher', () => {
+    it('identifies supported document file types', () => {
+      expect(isSupportedDocument('report.pdf')).toBe(true);
+      expect(isSupportedDocument('doc.docx')).toBe(true);
+      expect(isSupportedDocument('slides.pptx')).toBe(true);
+      expect(isSupportedDocument('data.xlsx')).toBe(true);
+      expect(isSupportedDocument('data.csv')).toBe(true);
+      expect(isSupportedDocument('image.png')).toBe(false);
+      expect(isSupportedDocument('audio.mp3')).toBe(false);
+    });
+
+    it('returns appropriate MIME types for supported documents', () => {
+      expect(getDocumentMimeType('report.pdf')).toBe('application/pdf');
+      expect(getDocumentMimeType('doc.docx')).toBe('application/vnd.openxmlformats-officedocument.wordprocessingml.document');
+      expect(getDocumentMimeType('slides.pptx')).toBe('application/vnd.openxmlformats-officedocument.presentationml.presentation');
+      expect(getDocumentMimeType('data.xlsx')).toBe('application/vnd.openxmlformats-officedocument.spreadsheetml.sheet');
+      expect(getDocumentMimeType('data.csv')).toBe('text/csv');
+      expect(getDocumentMimeType('unknown.xyz')).toBe('application/octet-stream');
+    });
+
+    it('dispatches extraction based on file extension', async () => {
+      const csvPath = path.join(tempDir, 'test.csv');
+      fs.writeFileSync(csvPath, 'Name,Role\nAlice,Developer\nBob,Designer');
+      const result = await extractDocument(csvPath);
+      expect(result.success).toBe(true);
+      expect(result.markdown).toContain('| Name | Role |');
+      expect(result.markdown).toContain('| Alice | Developer |');
+    });
+  });
 });