139_spanish-voice-trainer/main.py

# main.py
import asyncio
from docling.document_converter import DocumentConverter

async def dump_docling_output():
    file_path = "aula_int_plus_1_glos_en_alfa.pdf"
    output_path = "docling_output.md"
    
    print(f"🔄 Docling is analyzing structural layout for '{file_path}'...")
    
    try:
        converter = DocumentConverter()
        result = converter.convert(file_path)
        markdown_text = result.document.export_to_markdown()
        
        # Save the exact output to a file in your project root
        with open(output_path, "w", encoding="utf-8") as f:
            f.write(markdown_text)
            
        print(f"📝 Success! Raw Docling layout exported cleanly to: {output_path}")
        print("👀 Open this file in VS Code to see exactly how the text flows.")
        
    except Exception as e:
        print(f"❌ Docling processing failed: {e}")

if __name__ == "__main__":
    asyncio.run(dump_docling_output())