Movie Finder: semantic search over 1M films, inside MySQL

Movie Finder is a demo application built on MyVector, a vector-search component for MySQL, letting users semantically search up to 1,035,695 TMDB movies by describing them in natural language and getting results in milliseconds. It runs entirely inside MySQL 9.7 with no separate vector database, using HNSW indexing declared via a column comment and queried with functions like myvector_ann_set() and myvector_distance(). The demo shows filtering combined with vector search, 'more like this' queries using stored vectors, live index status, and speed-vs-recall comparisons between HNSW and exact scans. Benchmarks on a 16-core Arm host show HNSW search at 5-20ms versus 3 seconds for exact scan at full scale, with 100% recall@10 unfiltered. The app can be run locally via docker compose, with a live hosted demo also available.

Nguồn: https://askdba.net/movie-finder-semantic-search-over-1m-films-inside-mysql. 8 Sync News chỉ tóm tắt và dẫn link; bản quyền nội dung thuộc tác giả và nguồn gốc.