# Bengali OCR Web Application A modern, full-stack web application for extracting and editing Bengali text from scanned historical PDF documents. This is a React + FastAPI migration of the original CustomTkinter desktop application, maintaining 100% backward compatibility with the existing Python OCR logic. ## Features - **Split-screen interface**: PDF viewer on the left, text editor on the right - **PDF upload**: Drag-and-drop or file browser upload - **OCR processing**: - Local Tesseract OCR engine (primary) - Google Gemini API fallback - Batch processing for multiple pages - Real-time progress tracking - **Text editing**: Edit extracted Bengali text with auto-save - **Text export**: Export all pages to Word (.docx) document - **Settings management**: Configure Tesseract path, DPI, Gemini API key - **Book management**: View and switch between previously uploaded PDFs - **Bengali language support**: Full Bengali font rendering with proper diacritics ## Tech Stack ### Frontend - **React 18** with TypeScript - **Tailwind CSS** for styling - **Zustand** for state management - **Axios** for API communication - **Vite** as build tool ### Backend - **FastAPI** (Python web framework) - **SQLite** database - **PyMuPDF** for PDF handling - **Tesseract OCR** for local text recognition - **Google Generative AI** for Gemini API integration - **python-docx** for Word export ## Project Structure ``` . ├── frontend/ # React web application │ ├── src/ │ │ ├── components/ # Reusable React components │ │ ├── hooks/ # Custom React hooks │ │ ├── services/ # API client │ │ ├── store/ # Zustand state management │ │ ├── styles/ # Global styles & Tailwind │ │ ├── App.tsx # Main app component │ │ └── main.tsx # React entry point │ ├── index.html # HTML template │ ├── package.json # Dependencies │ ├── tsconfig.json # TypeScript config │ ├── vite.config.ts # Vite bundler config │ └── tailwind.config.ts # Tailwind configuration │ ├── backend/ # FastAPI server │ ├── app/ │ │ ├── routes/ # API endpoints │ │ ├── models.py # Pydantic schemas │ │ ├── config.py # Settings manager │ │ ├── main.py # FastAPI app │ │ ├── database.py # SQLite manager (unchanged) │ │ ├── ocr_processor.py # OCR pipeline (unchanged) │ │ └── bengali_ocr_gemini.py # Gemini integration (unchanged) │ ├── requirements.txt # Python dependencies │ └── run.py # Entry point script │ ├── config.json # Application settings ├── bengali_ocr.db # SQLite database └── tessdata/ └── ben.traineddata # Bengali language pack for Tesseract ``` ## Installation ### Prerequisites - Python 3.8+ - Node.js 16+ - Tesseract OCR (optional, for local processing) - Git ### Backend Setup ```bash # Navigate to backend directory cd backend # Create virtual environment python -m venv venv # Activate virtual environment # On Windows: venv\Scripts\activate # On Linux/Mac: source venv/bin/activate # Install dependencies pip install -r requirements.txt ``` ### Frontend Setup ```bash # Navigate to frontend directory cd frontend # Install dependencies npm install # Create .env file (optional, defaults provided) cp .env.example .env ``` ## Running the Application ### Start Backend ```bash cd backend python run.py # Backend will start at http://localhost:8000 # API docs available at http://localhost:8000/docs ``` ### Start Frontend (in another terminal) ```bash cd frontend npm run dev # Frontend will start at http://localhost:5173 # Automatically proxies /api requests to backend ``` ### Access the Application Open your browser and navigate to: `http://localhost:5173` ## Configuration Settings are stored in `config.json`: ```json { "tesseract_path": "C:/Program Files/Tesseract-OCR/tesseract.exe", "dpi": 200, "gemini_api_key": "your-api-key-here", "use_gemini_fallback": true, "db_path": "bengali_ocr.db" } ``` Or configure through the Settings modal in the UI. ## API Endpoints ### Books Management - `POST /api/books/upload` - Upload PDF file - `GET /api/books/` - List all books - `GET /api/books/{book_id}/progress` - Get processing progress - `DELETE /api/books/{book_id}` - Delete book ### Pages & OCR - `GET /api/pages/{book_id}/{page_num}` - Get page data and image - `POST /api/pages/{page_num}/ocr` - Process single page - `POST /api/pages/batch-ocr` - Start batch OCR processing - `GET /api/pages/task/{task_id}/status` - Check task status - `POST /api/pages/{page_num}/update-text` - Save edited text - `PUT /api/pages/{page_num}/reset` - Reset to original OCR text ### Configuration - `GET /api/config/` - Get current settings - `PUT /api/config/` - Update settings ### Export - `GET /api/export/books/{book_id}` - Export to Word format - `GET /api/export/books/{book_id}/preview` - Preview before export ## Usage Guide ### 1. Upload a PDF - Click "পিডিএফ ফাইল আপলোড করুন" button, or - Drag and drop a PDF file onto the application ### 2. Process Pages - **Single Page**: Select a page and click "এই পৃষ্ঠা প্রক্রিয়া করুন" - **All Pages**: Click "সব পৃষ্ঠা প্রক্রিয়া করুন" ### 3. Edit Text - Click on text in the right editor panel to edit - Changes are automatically saved - Use font size slider to adjust text size ### 4. View Other PDFs - Click "📚 আমার বইগুলি" to see all uploaded books - Click a book to switch to it ### 5. Configure Settings - Click "⚙️ সেটিংস" to configure: - Tesseract path - PDF DPI resolution - Gemini API key ### 6. Export Results - Click "📥 Word ফাইলে রপ্তানি করুন" to export all processed pages to .docx ## Testing ### Backend Testing ```bash # Run FastAPI in development mode with auto-reload cd backend python run.py # View API documentation # Visit: http://localhost:8000/docs ``` ### Frontend Testing ```bash # Run development server cd frontend npm run dev # Build for production npm run build # Preview production build npm run preview ``` ## Performance Notes - **DPI Impact**: Higher DPI (300) = better quality but slower processing - **Batch Processing**: Large batch sizes may take time; check progress bar - **Memory**: Keep browser DevTools closed for better performance - **Text Rendering**: Bengali fonts use system fonts with fallbacks for compatibility ## Troubleshooting ### Backend won't start - Ensure Python 3.8+ is installed - Check if port 8000 is available: `netstat -an | grep 8000` - Verify all dependencies: `pip install -r requirements.txt` ### Frontend shows blank page - Check browser console for errors (F12) - Ensure backend is running at `http://localhost:8000` - Clear browser cache and reload ### OCR not working - Verify Tesseract is installed and path is correct in settings - Check Gemini API key if using fallback - Try processing with lower DPI first ### Bengali text displays incorrectly - Ensure Bengali fonts are installed on system - Check browser font rendering settings - Try different browser (Chrome/Firefox) ## Database Schema The application uses SQLite with the following tables: ### Books Table ```sql CREATE TABLE books ( book_id INTEGER PRIMARY KEY, title TEXT NOT NULL, pdf_path TEXT NOT NULL, total_pages INTEGER NOT NULL, processed_pages INTEGER DEFAULT 0, upload_date TIMESTAMP DEFAULT CURRENT_TIMESTAMP ); ``` ### Pages Table ```sql CREATE TABLE pages ( page_id INTEGER PRIMARY KEY, book_id INTEGER NOT NULL, page_num INTEGER NOT NULL, raw_text TEXT, edited_text TEXT, page_image BLOB, status TEXT DEFAULT 'pending', error_message TEXT, FOREIGN KEY (book_id) REFERENCES books(book_id), UNIQUE(book_id, page_num) ); ``` ## Future Enhancements - [ ] WebSocket support for real-time progress updates - [ ] Multi-language OCR support - [ ] Document preprocessing optimization - [ ] Handwriting recognition - [ ] Batch import multiple PDFs - [ ] Collaborative editing - [ ] Cloud storage integration - [ ] Mobile-responsive interface ## License This project maintains the structure and dependencies of the original Bengali OCR application with modern web technologies. ## Support For issues or questions: 1. Check the troubleshooting section above 2. Review API docs at `http://localhost:8000/docs` 3. Check browser console for detailed errors 4. Verify backend logs for processing errors --- **Last Updated**: June 2026 **Version**: 1.0.0 # text_extractor