BRAINBYTE_SEARCH_ENGINE

Introduction

A free, open-source, and offline research engine based on BM25 and BERT models. Built using PyTorch and PyQt6.

Usage

If you want to download the complete project, please click the link below:

Download Complete Project

Note that you can only input queries listed in /data/sample_data/sample_queries.tsv.Because the complete dataset is too large, only test dataset was provided

If you only want to run the UI, you don't need to download /data/train/top100_train.csv.

Conda environment

First, install the Python dependencies:

conda create -n BRAINBYTE_SEARCH_ENGINE python==3.9
conda activate BRAINBYTE_SEARCH_ENGINE
pip install -r environment.txt

Setting PYTHONPATH in Conda Environment

Setting the PYTHONPATH environment variable after activating your environment allows Python to find and import modules from directories not installed in the standard library or virtual environment paths.

For Unix/Linux/macOS:

export PYTHONPATH="${PYTHONPATH}:/path/to/your/module"

For Windows:

set PYTHONPATH=%PYTHONPATH%;C:\path\to\your\module

To run the UI:

python search_enging_UI.py