yandexgpt-client

This library is written in Java 17 and provides a synchronous Java client for Yandex GPT completion API.

1.1 Prerequisites

Created an authorization api-key and folder ID for Yandex Cloud, official doc or here.

1.2 Configuration and Spring Boot dependency

You can leverage the yandexgpt-client library, for this clone the GitHub repository

git clone https://github.com/cheernota/yandexgpt-client

build the library:

mvn clean install

and add it as a dependency to your pom.xml:

<dependency>
    <groupId>com.cheernota</groupId>
    <artifactId>yandexgpt-client</artifactId>
    <version>1.0.0</version>
</dependency>

or build.gradle

implementation 'com.cheernota:yandexgpt-client:1.0.0'

Then add API-key and folder-ID to your application properties file

spring.ai.yandexgpt.folder-id=j39shh ...
spring.ai.yandexgpt.api-key=jkSS82kHH ...

1.3 Use case

After configuring you can autowire bean YandexGptChatClient and use its methods in your project like this:

private final YandexGptChatClient chatClient;

    @Autowired
    public GptService(YandexGptChatClient chatClient) {
        this.chatClient = chatClient;
    }
    
    public String askSimplePrompt() {
        String prompt = "Привет, расскажи насколько суров климат Сибири.";
        return chatClient.generateByPrompt(prompt);
    }

1.4 Other properties

Property	Description	Default
spring.ai.yandexgpt.base-url	Base url.	https://llm.api.cloud.yandex.net/foundationModels/v1
spring.ai.yandexgpt.enable-logging	Special header for managing the logging of your request data.	false

You can also configure the LLM request parameters manually, read the javadoc to the ChatCompletionRequest.