Bm25-se NPM | npm.io

bm25-se

This is a simple search engine implementation that can index and search documents based on the input query. The search engine uses a scoring algorithm based on the BM25 ranking function.

Installation

To use this search engine, you need to have Node.js installed on your machine. You can clone this repository and run npm install to install any dependencies required.

npm i bm25-se

Usage

const { SearchEnginee, TokenizerEn } = require('bm25-se')

// create a new search engine instance
const searchEngine = new SearchEnginee({ tokenizer: new TokenizerEn() })

// add documents to the index
searchEngine.addDocument("this is an example document")
searchEngine.addDocument("this is another document")
searchEngine.addDocument("yet document")

// search for documents that contain the word "example"
const searchResults = searchEngine.search("example document")

console.log(searchResults) // should output a list of documents that match the query, ordered by relevance

searchResults

Demo

You can try out the search engine in your browser by running the following command:

git clone https://github.com/Selenium39/bm25-se.git 
cd bm25-se
npm install
npm run web

Then open http://localhost:3000 in your browser.

demo

API

SearchEnginee(options)

Creates a new instance of the search engine.

options (optional): An object with the following properties:
- tokenizer (optional): A tokenizer object to use for tokenizing the documents. Defaults to new TokenizerEn().
- k1 (optional): A parameter for the BM25 scoring function. Defaults to 1.2.
- b (optional): A parameter for the BM25 scoring function. Defaults to 0.75.