Provide true fuzzy match options in Supabase JS Text Search

Currently, the textSearch functionality within Supabase JS is limited to options like "plain", "phrase", and "websearch". However, it lacks support for typo-correction or genuine fuzzy matching capabilities. The current implementation of .textSearch() and the underlying tsvector functionality relies on lexemes. However, it is worth noting that this approach (as described in the [docs]( is presented as "fuzzy matching”, but might not entirely align with the conventional understanding of fuzziness. Encountering limitations when dealing with mispelled non-standard English words or those without a corresponding lexeme can be a bit frustrating for users - depending on the domain of the application. I am wondering if there are any plans to introduce some form of fuzzy matching feature within the JavaScript client. This may potentially need to utilise extensions such as pg_bm25, fuzzystrmatch, pg_trgm, or work by making adjustments to the to_tsvect…