Skip to content

This is a tool to compare two different Elasticsearch or OpenSearch queries against each other to help evaluate the relevance of the results returned.

Notifications You must be signed in to change notification settings

tylercrosse/relevance-tuning-tool

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

23 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Relevance Tuning Tool

Screenshot of the top portion of the tool Screenshot of the bottom portion of the tool

Background

This is a tool to compare two different Elasticsearch or OpenSearch queries against each other to evaluate the relevance of the results returned. This project is inspired by the search comparison tool plugin, that was recently released to OpenSearch in 2.4.0. Here's a link to the repo for that tool, which uses the osd-manaco package from the main OpenSearch-Dashboards repo.

The OpenSearch cluster I have access to is still running OpenSearch v1.3 and I wanted to play around with the relevance of the search results in a few indexes I work on. I thought this would be a fun project to spin up. I wanted to help improve the relevance of our search results after recently reading Relevant Search by Doug Turnbull and John Berryman, as well as AI-Powered Search by Tray Grainger, Doug Turnbull, and Max Irwin.

Architecture

Front end

  • Create React App
  • Ace Editor
  • React Table

Back end (proxy server)

  • Flask
  • AWS SDK

Getting Started

Auth

This project currently assumes you have an ES or OpenSearch instance running on AWS. I'd like to support more auth in the future. It uses requests-aws4auth to read your aws_access_key_id and aws_secret_access_key from your default profile in your ~/.aws/credentials file.

Dependencies

  1. Install the node modules

npm install

  1. Install the python3 dependencies

cd api && pip install -r requirements.txt

Start the front end and API

npm run start kicks off the front end and should automatically open your default browser at http://localhost:3000 python api/api.py will start the API. You will need to run this in a new tab.

About

This is a tool to compare two different Elasticsearch or OpenSearch queries against each other to help evaluate the relevance of the results returned.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published