Book Image

Redis Stack for Application Modernization

By : Luigi Fugaro, Mirko Ortensi
1 (1)
Book Image

Redis Stack for Application Modernization

1 (1)
By: Luigi Fugaro, Mirko Ortensi

Overview of this book

In modern applications, efficiency in both operational and analytical aspects is paramount, demanding predictable performance across varied workloads. This book introduces you to Redis Stack, an extension of Redis and guides you through its broad data modeling capabilities. With practical examples of real-time queries and searches, you’ll explore Redis Stack’s new approach to providing a rich data modeling experience all within the same database server. You’ll learn how to model and search your data in the JSON and hash data types and work with features such as vector similarity search, which adds semantic search capabilities to your applications to search for similar texts, images, or audio files. The book also shows you how to use the probabilistic Bloom filters to efficiently resolve recurrent big data problems. As you uncover the strengths of Redis Stack as a data platform, you’ll explore use cases for managing database events and leveraging introduce stream processing features. Finally, you’ll see how Redis Stack seamlessly integrates into microservices architectures, completing the picture. By the end of this book, you’ll be equipped with best practices for administering and managing the server, ensuring scalability, high availability, data integrity, stored functions, and more.
Table of Contents (18 chapters)
1
Part 1: Introduction to Redis Stack
6
Part 2: Data Modeling
11
Part 3: From Development to Production

Performing hybrid queries

Hybrid queries are VSS queries mixed with ordinary search algorithms (numeric, text, tag, and geo). When running hybrid queries with VSS, it is possible to include business logic in the query to enrich the search criteria and simplify the client application code. These conventional filters are pre-filters to the vector search operation and are meant to simplify the similarity search by reducing the computational effort to retrieve the KNN results. An example based on the previous proof of concept can be written by replacing * with the desired query. In this case, this is a filter with the genre tag that retrieves the closest documents in the “technical” category:

q = Query("@genre:{technical}=>[KNN 2 @embedding $vec AS score]")
               .return_field("score")
             ...