Exploring Spring AI 1.0.0 M2: A Hands-On Guide with Code Samples

Welcome to our deep dive into Spring AI 1.0.0 M2, the latest release designed to simplify AI integration within the Spring framework. This update brings a host of enhancements that streamline the process of incorporating AI functionalities into your applications. In this post, we'll explore these new features and walk through a practical example of using Spring AI 1.0.0 M2 with a TensorFlow model.

What's New in Spring AI 1.0.0 M2?

Spring AI 1.0.0 M2 introduces several exciting updates:

• Simplified Configuration: Less boilerplate code and easier setup for AI components.
• Expanded AI Framework Support: Enhanced compatibility with TensorFlow, PyTorch, and other popular AI frameworks.
• Performance Enhancements: Improved efficiency and scalability for AI applications.
• Improved Documentation: Updated guides and examples for quicker implementation.

Getting Started with Spring AI 1.0.0 M2

To showcase the capabilities of Spring AI 1.0.0 M2, we'll create a simple Spring Boot application that uses TensorFlow for text classification. Follow these steps to get started:

1. Create a New Spring Boot Project

Generate a new Spring Boot project using Spring Initializr. Add the following dependencies:

• Spring Web
• Spring Boot DevTools
• Spring Data JPA (optional)

2. Add TensorFlow Dependency

Update your pom.xml file with the TensorFlow dependency:

<dependencies>
    <!-- Spring Boot Starter Web -->
    <dependency>
        <groupId>org.springframework.boot</groupId>
        <artifactId>spring-boot-starter-web</artifactId>
    </dependency>
    
    <!-- TensorFlow Java Library -->
    <dependency>
        <groupId>org.tensorflow</groupId>
        <artifactId>tensorflow</artifactId>
        <version>2.11.0</version>
    </dependency>
</dependencies>

3. Build the AI Component

Create a service class to handle interactions with the TensorFlow model:

package com.example.springai.service;

import org.springframework.stereotype.Service;
import org.tensorflow.SavedModelBundle;
import org.tensorflow.Tensor;

@Service
public class TextClassificationService {

    private final SavedModelBundle model;

    public TextClassificationService() {
        // Load the pre-trained TensorFlow model
        this.model = SavedModelBundle.load("path/to/saved_model", "serve");
    }

    public String classifyText(String text) {
        try (Tensor<String> inputTensor = Tensor.create(text.getBytes("UTF-8"), String.class)) {
            // Run inference
            Tensor<?> result = model.session().runner()
                .feed("input_tensor_name", inputTensor)
                .fetch("output_tensor_name")
                .run()
                .get(0);

            // Process result
            float[][] output = result.copyTo(new float[1][1]);
            return output[0][0] > 0.5 ? "Positive" : "Negative";
        } catch (Exception e) {
            throw new RuntimeException("Error during text classification", e);
        }
    }
}

Note: Replace "path/to/saved_model", "input_tensor_name", and "output_tensor_name" with the actual paths and names used in your TensorFlow model.

4. Create a Controller

Add a REST controller to expose an API endpoint for text classification:

package com.example.springai.controller;

import com.example.springai.service.TextClassificationService;
import org.springframework.web.bind.annotation.GetMapping;
import org.springframework.web.bind.annotation.RequestParam;
import org.springframework.web.bind.annotation.RestController;

@RestController
public class TextClassificationController {

    private final TextClassificationService classificationService;

    public TextClassificationController(TextClassificationService classificationService) {
        this.classificationService = classificationService;
    }

    @GetMapping("/classify")
    public String classify(@RequestParam String text) {
        return classificationService.classifyText(text);
    }
}

Running the Application

1. Build and Run Your Application

   Use Maven to build and run your Spring Boot application:

   mvn clean install
   mvn spring-boot:run
   

2. Test the API

   Open your browser or use a tool like curl or Postman to test the endpoint:

   
   curl "http://localhost:8080/classify?text=I%20love%20Spring%20AI"
   

   You should receive a response indicating whether the text is classified as "Positive" or "Negative."