Skip to content

Latest commit

 

History

History
71 lines (57 loc) · 1.97 KB

README.md

File metadata and controls

71 lines (57 loc) · 1.97 KB

Doc Retrieval

image image image

With this app, you can upload your PDF file and search for a specific words, phrase and etc. The main algorithm can describe bellow:

image

Deployment

To deploy this project run

1: Clone this repository:

git clone [email protected]:shirindehghani/Information-Retrieval.git

2: install requirements

pip install -r requirements.txt

Docker

Run bellow command to make docker image:

docker build -t your_image_name .

You can also use this app base on streamlit UI command:

streamlit run main.py

You should import you OpenAI api_key, input query and a PDF file:

image

Project Structure

|──IR_models ──|──Retrieval.py
|──configs ──|──configs.json
|──imgs ──|──algorithm.png
|         |──Ui.png
|──temp ──|──PDF-file.pdf
|──needded_pkgs ──|──Log-Handler
|──app.py
|──docker-compose.yml
|──Dockerfile
|──LICENCE
|──main.py
|──main2.py (streamlit)
|──README.md
|──requirements.txt
|──test_app.py

TODO:

  • Improve the UI
  • Speed up the process and improve chuncks

References


@article{zhu2023large,
  title={Large language models for information retrieval: A survey},
  author={Zhu, Yutao and Yuan, Huaying and Wang, Shuting and Liu, Jiongnan and Liu, Wenhan and Deng, Chenlong and Dou, Zhicheng and Wen, Ji-Rong},
  journal={arXiv preprint arXiv:2308.07107},
  year={2023}
}