Marker-Inc-Korea / AutoRAG

AutoML tool for RAG
https://auto-rag.com/
Apache License 2.0
2.7k stars 211 forks source link

[Feature Request] Add Windows Support #762

Closed vkehfdl1 closed 1 month ago

vkehfdl1 commented 1 month ago

Is your feature request related to a problem? Please describe. Some people still using windows. We need to test AutoRAG on the Windows machine.

Describe the solution you'd like Test the pytest first. And then if something’s wrong, we will fix it.

Describe alternatives you've considered I think we need to analyze how many people downloading AutoRAG in Windows.

Additional context Add any other context or screenshots about the feature request here.

vkehfdl1 commented 1 month ago

https://docs.python.org/ko/3/library/tempfile.html https://www.scivision.dev/python-tempfile-permission-error-windows/

Most of the error occured because of the tempfile library!!

vkehfdl1 commented 1 month ago

TART will not supported on the Windows. (tokenizer byte sequence problem)

Plus, you have to install this.

import nltk
nltk.download('averaged_perceptron_tagger_eng')

Plus, I got trouble to install libmagic. Install libmagic at Windows is tricky....

We need to change libmagic to the puremagic for removing OS dependency. Why windows libmagic is not working properly? And it turns out unstructured using libmagic.

https://docs.unstructured.io/open-source/installation/full-installation

And they are forced me to use docker images for that only.

vkehfdl1 commented 1 month ago

Also UPR will not supported on the Windows, with the same reason. MonoT5 as well.

vkehfdl1 commented 1 month ago

It looks like it is hard to support konlpy korean tokenizers on Windows. Hard to install it.