server.py

import os
import base64
from marker.convert import convert_single_pdf
from marker.models import load_all_models
from marker.settings import Settings
import gradio as gr


model_list = load_all_models()

def parse_pdf_and_return_markdown(pdf_file: bytes , extract_images: bool):
    full_text, images, out_meta = convert_single_pdf(pdf_file, model_list)
    image_data = {}
    if extract_images:
        for i, (filename, image) in enumerate(images.items()):
            image_filepath = f"image_{i+1}.png"
            image.save(image_filepath, "PNG")

            with open(image_filepath, "rb") as f:
                image_bytes = f.read()

            image_base64 = base64.b64encode(image_bytes).decode('utf-8')
            image_data[f'image_{i+1}'] = image_base64

            os.remove(image_filepath)

    return full_text, out_meta, image_data
    

# @app.post("/convert")
# async def convert_pdf_to_markdown(pdf_file: UploadFile = File(...), extract_images: bool = True):
#     if extract_images == False:
#         Settings.EXTRACT_IMAGES = False
#         print("Print EXTRACT_IMAGES set to False")
#     else:
#         Settings.EXTRACT_IMAGES = True
#     if pdf_file.content_type != "application/pdf":
#         raise HTTPException(
#             status_code=status.HTTP_415_UNSUPPORTED_MEDIA_TYPE,
#             detail=f'File {pdf_file.filename} has unsupported extension type',
#         )
#     markdown_text, metadata, image_data = parse_pdf_and_return_markdown(await pdf_file.read(), extract_images=extract_images)
#     return {"markdown": markdown_text, "metadata": metadata, "images": image_data }

# def main():
#     # Parse command-line arguments
#     parser = argparse.ArgumentParser(description="Run the marker-api server.")
#     parser.add_argument("--host", default="127.0.0.1", help="Host IP address")
#     parser.add_argument("--port", type=int, default=8000, help="Port number")
#     args = parser.parse_args()

#     # Load all models before starting the server
#     configure_logging()  # Assuming this function initializes logging    

#     # Start the server
#     import uvicorn
#     uvicorn.run(app, host=args.host, port=args.port)

with gr.Blocks() as server:
    gr.Markdown("Upload a PDF file to convert to markdown.")
    gr.Interface(
        parse_pdf_and_return_markdown, 
        inputs=[gr.File(label="Upload PDF", type="filepath"), gr.Checkbox(label="Extract Images")],
        outputs=[gr.Textbox(label="Markdown"), gr.JSON(label="Metadata"), gr.JSON(label="Images")]
    )


if __name__ == "__main__":
    server.launch()