run.py

import sys
import argparse
import logging
from pathlib import Path

# Setup logging
logging.basicConfig(level=logging.INFO, format="%(asctime)s - %(levelname)s - %(message)s")
log = logging.getLogger("OCR_Runner")

def main():
    parser = argparse.ArgumentParser(description="Clean & Public OCR Repository Runner")
    parser.add_argument("input", help="Path to the input PDF file")
    parser.add_argument("--engine", choices=["docling", "paddle", "hybrid"], default="docling", 
                        help="OCR engine to use (default: docling)")
    parser.add_argument("--output", default="output", help="Output directory (default: output)")
    parser.add_argument("--gpu", action="store_true", default=True, help="Use GPU if available (default: True)")
    parser.add_argument("--no-gpu", action="store_false", dest="gpu", help="Force CPU usage")
    
    args = parser.parse_args()
    
    input_path = Path(args.input)
    if not input_path.exists():
        log.error(f"❌ Input file not found: {args.input}")
        sys.exit(1)

    log.info(f"🚀 Starting OCR with {args.engine} engine...")

    try:
        if args.engine == "docling":
            from engines.docling_engine import DoclingEngine
            engine = DoclingEngine(use_gpu=args.gpu)
        elif args.engine == "paddle":
            from engines.paddle_engine import PaddleEngine
            engine = PaddleEngine(use_gpu=args.gpu)
        elif args.engine == "hybrid":
            from engines.hybrid_engine import HybridEngine
            engine = HybridEngine(use_gpu=args.gpu)
        
        text_path, meta_path = engine.run(str(input_path), output_dir=args.output)
        
        log.info("✅ OCR Process Completed Successfully!")
        log.info(f"📄 Text saved to: {text_path}")
        if meta_path:
            log.info(f"🧠 Metadata saved to: {meta_path}")

    except Exception as e:
        log.error(f"❌ An error occurred: {e}")
        import traceback
        traceback.print_exc()
        sys.exit(1)

if __name__ == "__main__":
    main()