awesome-search-relevance

Tools and other things for people who work on search relevance & information retrieval

Tools

The "plungers" we use in our job day-to-day.

Offline Search Relevance Testing

Quepid - test driven relevance browser sandbox using a small set of judgments
Splainer - debug a single Solr or Elasticsearch query
Rated Ranking Evaluator - Maven based relevance testing tool against a Solr or Elasticsearch config
Quaerite - Search relevance evaluation framework

Search Engines

Learning to Rank Tooling

Tooling for doing machine learning to optimize relevance

Search Engine Plugins

Elasticsearch Learning to Rank - supports storing and logging features (ES queries), uploading models, and executing Ranklib and XGBoost models
Solr Learning to Rank - supports storing and logging features (Solr queries), uploading models, and executing a range of models

Learning to Rank Training Models

Ranklib - original implementation of LambdaMART and other algorithms
RankyMcRankFace - Ranklib fork with a few extra features
XGBoost - eXtreme Gradient Boosting - scalably train a MART ranking model
LightGBM - Microsoft's gradient boosting library
TFRanking - Tensorflow ranking

NLP

Gensim - Topic modeling, and other useful NLP tasks like collocation detection, word2vec, and others

Analytics

Search Collector - Collecting behavioral signals from a browser

Rule Systems

Querqy - Query preprocessing & search rule system for applying query-specific boosts, filters, downboosts, etc (ie downboost 'case' category when user searches for 'iPhone')
SMUI - UI for managing Solr collections, which supports (among other things) managing Quergy rules.

Blogs

Lucidworks Blog - Fusion & Solr
OpenSource Connections Blog - open source relevance blogging

Conferences & Meetups

Haystack - Haystack the Search Relevance conference
Activate - formerly Lucene/Solr Revolution, focus on Solr+AI
SIGIR - ACM's Information Retrieval conference. Academic focused with an industry day
ECIR - European Conference on Information Retrieval
MICES - MIx Camp E-commerce Search, focused on E-commerce industry across technologists and product managers
Search Solutions - British Computer Society day covering search and information retrieval

Past Conference Videos

Haystack US 2019 - Youtube Playlist of Haystack 2019 talks.

Training

OpenSource Connections Training - Solr & Elasticsearch relevance, Learning to rank, NLP training and others
Elastic training - a smattering of classes that can help with relevance like deep diving into analyzers, etc
Lucidworks training - classes covering Solr + Lucidworks Fusion

Hosted Cloud Based Open Source Search

Cloud-based Solr, Elasticsearch