BUG FIXES: 1. EasyOCR 'Invalid input type' error: - PIL Images must be converted to numpy arrays - Added np.array() conversion before passing to EasyOCR 2. PaddleOCR 'Unknown argument: show_log' error: - Some versions don't support show_log parameter - Added try/except to handle both versions - Falls back to creating PaddleOCR without show_log NEW FEATURES: 1. OCR Backend Manager (core/ocr_backend_manager.py): - Detects installed OCR backends - Checks Windows Registry for Tesseract installation - Searches common installation paths - Auto-configures pytesseract with found binary 2. Calibration Tab Improvements: - Added 'Install EasyOCR' button (pip install easyocr) - Added 'Install Tesseract Package' button (pip install pytesseract) - Added 'Auto-Detect Tesseract' button (checks registry/paths) - Added 'Install PaddleOCR' button (pip install paddleocr) - Shows which backends are available and where they're installed 3. Better Error Messages: - Clear instructions when backends are missing - Direct links to download pages - Shows detected installation paths The app now: - Automatically finds Tesseract from registry - Allows installing backends from within the UI - Properly handles all OCR input formats |
||
|---|---|---|
| .. | ||
| __init__.py | ||
| plugin.py | ||