The ReadMe did not tell how to run it.

[yutaoxie]

Sorry I could not follow your Readme...

cd MinerU --- git clone will create the MinerU directory... so do you want us to clone first?

cd model ---- also confusing.

It might be helpful to tell the final dictory stucture.

Step 6: where is the magic_pdf.json?

And finally, how to run this tool? command line? webUI?

[jxtse]

Hello Yutao,

Thank you for raising this issue!

Yes, this project depends on the MinerU repository, so you'll need to clone it first and follow their instructions to download magic_pdf.json.

The final directory structure should be as follows:

PaperToSlides/
├── data/ # Example data
│ ├── Example.pdf
├── MinerU/ # MinerU source code
│ ├── demo/
│ │ ├── ...
│ ├── magic_pdf/
│ │ ├── model
│ │ ├── ...
│ ├── ...
├── output/ # Example output
├── GenerateSlidesOutline.py
├── GenerateSlidesOutlinePrompt.md
├── MinerU.py
├── .gitignore
├── README.md

The model file is stored in C:\{User}\.cache\modelscope\hub\models\opendatalab\PDF-Extract-Kit and the magic_pdf.json is stored in C:\{User}\magic-pdf.json by default.

To run this tool, simply execute GenerateSlidesOutline.py. You need to input your "OPENAI_API_KEY" before executing it.

I understand that the current setup is a bit complex, and I plan to improve it in future updates. Thank you for your patience!

[yutaoxie]

Thanks, I was able to get it up and running. By default GenerateSlidesOutline.py did not use MinerU at all:
# md_file_path = extract_pdf_to_markdown(pdf_file_name, local_md_dir

Testing with MinerU.py, and now I run into another issue:

from magic_pdf.libs.MakeContentConfig import DropMode, MakeMode
from magic_pdf.pipe.OCRPipe import OCRPipe

There is no more MakeContentConfig or pipe.OCRPipe in the MinerU package. And the MinerU offical depot(https://github.com/opendatalab/MinerU/tree/master/magic_pdf) does not have "pipe" either.

[yutaoxie]

Thanks, I was able to get it up and running. By default GenerateSlidesOutline.py did not use MinerU at all: # md_file_path = extract_pdf_to_markdown(pdf_file_name, local_md_dir

Testing with MinerU.py, and now I run into another issue:

from magic_pdf.libs.MakeContentConfig import DropMode, MakeMode from magic_pdf.pipe.OCRPipe import OCRPipe

There is no more MakeContentConfig or pipe.OCRPipe in the MinerU package. And the MinerU offical depot(https://github.com/opendatalab/MinerU/tree/master/magic_pdf) does not have "pipe" either.

Well, I figured it out... I guess the current code is outdated. I found a working version here:

https://mineru.readthedocs.io/en/latest/user_guide/usage/api.html

This code works for me. A happy day.

[jxtse]

Glad to hear you got it working!

Thanks for pointing out the issue with the outdated code. If you have the time, feel free to submit a pull request to update this repository with the working version. That would be greatly appreciated!