{"id":69651,"date":"2025-06-20T09:43:20","date_gmt":"2025-06-20T01:43:20","guid":{"rendered":"https:\/\/hongtronics.com\/?p=69651"},"modified":"2025-06-20T09:53:53","modified_gmt":"2025-06-20T01:53:53","slug":"redis-springai-rag-solution-com","status":"publish","type":"post","link":"https:\/\/hongtronics.com\/en\/redis-springai-rag-solution-com\/","title":{"rendered":"Hong Hong Sharing | Building a Rag Application with Redis and Spring Ai"},"content":{"rendered":"
\n\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t

summaries<\/h2>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t
\n\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t

This article introduces Spring AI.Supporting databases are key storage technologies in LLM-driven applications<\/strong><\/span>Previously, Spring had few options for launching AI applications built with Redis, and the Spring community is working on a Spring AI project to simplify development. The Search Augmented Generation (RAG) technology combines data and AI models, using the beer dataset as an example.Describes how loading data into Redis implements the RAG workflow.<\/strong><\/span>The code is available on GitHub and relies on Spring Boot, Azure OpenAI, etc. It also introduces the classes responsible for data loading, RAG processes, packaging HTTP endpoints, and the React interface to interact with Spring endpoints, making it easier for Java developers to build AI applications.<\/span><\/span><\/p><\/div>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t

\n\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t\t\t\t
\n\t\t\t\t\t\t\t\t\t\t\"Redis\u6578\u64da\u6d41\u53ef\u8996\u5316\"\t\t\t\t\t\t\t\t\t\t\t
<\/figcaption>\n\t\t\t\t\t\t\t\t\t\t<\/figure>\n\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t
\n\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t

Introduction to Spring AI<\/h2>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t
\n\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t

In Large Language Model (LLM)-driven applications, the provision of databases often serves as the core storage technology for Artificial Intelligence applications. Such databases need to support semantic search and provide the relevant contextual environment for LLM.<\/span><\/p>

Until now, the options for building AI applications with Spring and Redis were relatively limited. Recently, Redis has attracted a lot of attention as a high-performance supporting database solution. The Spring community has launched a new project called Spring AI, which aims to streamline the development process for artificial intelligence applications, especially those involving supporting databases.<\/strong><\/span><\/span><\/p>

The following section describes how to build a Spring AI application that implements a search augmentation generation (RAG) workflow using Redis as the backing database.<\/span><\/p><\/div>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t

\n\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t

Search Enhanced Generation<\/h2>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t
\n\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t

Retrieval Augmented Generation (RAG) is a technical approach that combines data with artificial intelligence models.<\/strong><\/span>In a RAG workflow, data is first loaded into a supporting database (e.g. Redis). Upon receiving a user query, the supporting database will retrieve a set of documents that are similar to the query. These documents are used as the context for answering the user's question and are combined with the user's query to generate a response, usually through an artificial intelligence model.<\/span><\/p>

In this example, we will demonstrate this using a dataset containing information about various types of beers, with attributes such as the name of the beer, the alcohol content (ABV), the International Bitterness Units (IBU), and the description.<\/span><\/strong>The dataset will be loaded into Redis to demonstrate a practical application of the RAG workflow.<\/span><\/p><\/div>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t

\n\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t

Codes and Dependencies<\/h2>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t
\n\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t

It can be found on GitHubSpring AI and Redis<\/a>All the code for the demo.<\/span><\/p>

This project uses Spring Boot as a startup dependency for web applications, and combines Azure OpenAI and Spring AI Redis.<\/span><\/strong><\/span><\/p><\/div>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t

\n\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t

Data Loading<\/h2>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t
\n\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t

Our application will use JSON files that provide beer information as a data source. The structure of each file is as follows:<\/span><\/p><\/div>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t

\n\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t
{
\u201cid\u201d: \u201c00gkb9\u201d.,
\u201cname\u201d: \u201cSmoked Porter Ale\u201d.,
\u201cdescription\u201d: \u201cThe Porter Pounder Smoked Porter is a dark rich flavored ale that is made with 5 malts that include smoked and chocolate roasted malts. It has coffee and mocha notes that create a long finish that ends clean with the use of just a bit of dry hopping \u201d,
\u201cabv\u201d: 8.,
\u201cibu\u201d: 36
}<\/span><\/span><\/div><\/div>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t
\n\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t

To load the beer dataset into Redis, we will use the RagDataLoader class. This class contains a method that runs when the application starts. In this method, we use a JsonReader to parse the dataset and then insert the file into Redis using an automatically connected VectorStore.<\/span><\/p><\/div>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t

\n\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t
\/\/ Create a JSON reader with fields relevant to our use case JsonReader loader = new JsonReader(file, \u201cname\u201d, \u201c abv\u201d, \u201cibu\u201d, \u201cdescription\u201d); \/\/ Use the autowired VectorStore to insert the documents into Redis vectorStore.add(loader.get()); \/\/ Use the autowired VectorStore to insert the documents into Redis vectorStore.;<\/span><\/span><\/div><\/div>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t
\n\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t

At this point, we have obtained a dataset of about 22,000 beers and their corresponding embeddings.<\/span><\/p><\/div>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t

\n\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t

RAG Services<\/h2>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t
\n\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t

The RagService category implements the RAG workflow. When prompted by the user, it calls the retrieval method and performs the following steps:<\/span><\/p>