とりあえずローカルで全文検索をN-gramにするのはちょっと設定追記するだけで出来そう / It seems Fulltext search methods on Chewy/Elasticsearch are easily changeable.(Note: N-gram search method on PostgreSQL has some difficulty...) Feels good to me. I'd like to use N-gram tokenizer for fulltext search because default tokenizer may not work with Japanese non-space-separated texts.