retriever-search 0.5.1

Local retriever search for your use

Retriever

To setup your own local search you can now use this repo.

---

Install

Recommended to use a virtual environment using venv -

$ python -m venv new_env
$ source new_env/bin/activate

For installing from the github repo -

$ pip install -e .

For installing from the pip -

$ pip install retriever-search

Search server

>>> from retriever_search import search_server
>>> search_server.run_search_server('input_directory', 'input_json', 'json_save_path', 'embedding_model', 'qa_model', device='cpu')

Search parameter meanings

• input_directory -- The directory holding your files
• input_json -- pre saved json file from earlier runs can be used for faster loading
• json_save_path -- pass for saving the embeddings to a json can be used later as input_json
• embedding_model -- pick the embedding model you want to we use Spectre model as a default
• qa_model -- you can currently pick between tiny, medium and large

Frontend and Vizualisation run

>>> from retriever_search import frontend_app as fp
>>> fp.run_frontend()

Where to access the frontend

Access via the following URL - http://127.0.0.1:7860 This URL would work for your local setup only

Tickets

1.0.0

• Add lightweight embedding model support as flag in launcher
• Remove HTML requirement for plotly visualization
• Make LDA visualization update
• Add model quantized