Using Trigram pg extension for more performant text similarity searches

[maindsh]

It is a known fact that using LIKE and ILIKE queries all of rows are scanned in the table. Compared to using Trigram (pg extension) instead of sequentially scanning all of the rows in the document, only the gin index is scanned which is at least 100X faster.

https://mazeez.dev/posts/pg-trgm-similarity-search-and-fast-like

Related to this is Ransack *_matches (and others) which internally uses these LIKE queries https://activerecord-hackery.github.io/ransack/getting-started/search-matches/

So my question is, is there a way to override the Ransack matchers to use Similarity function similarity('your-search-term') instead of LIKE ?

[iamonkara]

+100

We are in much need of a performant full text search and I would be happy to implement it if some maintainer can share some clues on which classes to look into.