Building Real-Time Full-Text Site Search with RediSearch

We built a search experience for the Redis documentation site using RediSearch, which provided robust full-text search features, allowing us to transform a bland form into an awesome search experience. We created a Python application called redis-sitesearch that indexes our documentation site in RediSearch and powers a REST API backend that queries RediSearch. The frontend makes search requests to the backend API while rendering results. We run the pieces of the project inside a container on Google Cloud, distribute instances worldwide using a multi-zone deployment behind a global load balancer, and use a background job to scrape our documentation site and create Hashes for RediSearch to index. We validate documents based on arbitrary logic, score documents with fine-grained control, and use prefix matching for effective search queries. The application is deployed across multiple zones in Google Cloud, reducing latency and providing an excellent search experience.