0.9.0

• OpenAIFileAPI for File operation
• OpenAIBatchAPI for Batch operation

0.8.0

Structured Output

Please refer OpenAI Structured Outputs for detail.

First you need to define an record for structured output:

@StructuredOutput(name = "java_example")
public record JavaExample(@Parameter("explanation") String explanation, @Parameter("answer") String answer,
		@Parameter("code") String code, @Parameter("dependencies") List<String> dependencies) {
}

Then you can use structured output record as return type as following:

        @ChatCompletion(system = "You are a helpful Java language assistant.")
   	Mono<JavaExample> generateJavaExample(String question);

Attention: if the return type is not Mono<String>, and it means structured output.

0.7.0

Guys, sorry for long time to update.

• Models updated: chat(gpt-4o-mini), complete(gpt-3.5-turbo-instruct)
• Function call: use new tools instead of function_call(deprecated)

0.6.0

• Add embeddings and completion support
• Refactor: move chat related classes to chat package.

0.5.1

@joshlong did lots of contribution for this release, and thanks.

• Document updated by Josh Long
• Project structure adjusted by Josh Long, and Spring friendly now
• More powerful function resolution by Josh Long
• Improved function resolution in graalvm by Josh Long
• Prompt value could be from classpath and URL: conversation=classpath:///conversation-prompt.txt or conversation=https://example.com/conversation-prompt.txt
• Prompt value resolve with records support

0.4.0

• Prompt management added
• Prompt as Lambda/Function
• Reactive GPTFunction with Mono return type

0.3.0

ChatGPT interface support with functions built-in.

@GPTExchange
public interface GPTHelloService {

    @ChatCompletion("You are a language translator, please translate the below text to Chinese.\n")
    Mono<String> translateIntoChinese(String text);

    @ChatCompletion("You are a language translator, please translate the below text from {0} to {1}.\n {2}")
    Mono<String> translate(String sourceLanguage, String targetLanguage, String text);
}

Create ChatGPT interface service bean:

    @Bean
    public GPTHelloService gptHelloService(ChatGPTServiceProxyFactory proxyFactory) {
        return proxyFactory.createClient(GPTHelloService.class);
    }

0.2.1

• Add proxy() API in OpenAIChatAPI.java for OpenAI REST API proxy.
• Add OpenAIProxyController for proxy support.
• Add PromptManager demo based on properties file

0.2.0

• GraalVM native image support by @joshlong
• Introduce ChatRequestBuilder to build ChatGPT request
• Introduce ChatMessageRole enum

0.1.1

Initial release.

• Base on Spring Boot 3.0+
• Async with Spring Webflux
• Support ChatGPT Chat Stream
• Support ChatGPT functions