Join GitHub today
GitHub is home to over 36 million developers working together to host and review code, manage projects, and build software together.Sign up
ElasticSearch improvements (scoring, boosting, misspelling/fuzziness) #2340
Short description and why it's useful
There's currently no way to customize which fields to search with, neither the weights (boost) for each field, nor fuzziness (misspelling) or custom scoring functions . The default query uses a "match_phrase_prefix" operator which is usually used as autocomplete-like feature, but lacks of some key Elastic key features like: fuzziness, custom scoring, misspelling (for instance if your query term is "Hodie" instead of "Hoodie", you'll get no results), etc.
We have made a few modifications so all these features can be customized via configuration file, giving the ability to make the search query more dynamic and powerful than the current functionality. It is now possible to define the searchable fields, how to sort them by attribute values, added fuzziness (for misspelling) an ability to add custom weights (boosts) to both function_score and attributes individually.
Screenshots of visual changes before/after (if there are any)
Upgrade Notes and Changelog
Contribution and currently important rules acceptance