Unconference Session: We Need Query Rewriting
summary
Argues that OpenSearch lacks a business-facing query rewriting layer and shows Quirky, an open source rules engine used at massive scale (Walmart Labs) to rewrite queries into synonyms, category boosts, redirects and filters, avoiding global synonym side effects. Demonstrates fixing a bad “laptop” query with a per-query rule, publishing live, and improving results. Notes Quirky exists for Solr/Elasticsearch and an OpenSearch port at version 2.19.2, and describes dozens to thousands of human rules combined with ML generation and up/down/redirect/filter primitives.
AI-written summary · about
chapters (10)
- 0:00Query rewriting need
- 1:03Query rewriting for merchandisers
- 2:06Fixing laptop query demo
- 4:46Introducing Quirky tool
- 7:12Quirky use cases and benefits
- 9:18Quirky challenges in OpenSearch
- 10:45Machine learning with Quirky
- 11:47OpenSearch dashboard integration
- 13:15Multiple rules and validation
- 14:18Hurricane redirect example
related talks
09mUnconference: OpenSearch is not a database!
24mReimagining Search : Scaling OpenSearch Across an Enterprise - Platform C...
28mSuper Search with OpenSearch and Python
41mDynamoDB, Neptune and OpenSearch: A Unified Approach to Collaborative Editing and Search #JSConfAM25
05mKeynote: OpenSearch: From Code to Community
06mSponsored Keynote: Foundation-Led Innovation: OpenSearch's Impact on Modern Data I...
45mAnything New in Search? What If I Find It?
39mMastering OpenSearch Benchmark: Advanced Techniques for Power Users