WF-ChatGptApi:

Maven:

Java(min): 17

  <dependency>
    <groupId>io.github.wf4java</groupId>
    <artifactId>WF-ChatGptApi</artifactId>
    <version>1.1</version>
  </dependency>

Examples:

Create and generate:

ChatGpt gpt = new ChatGpt("api-key");

String result1 = gpt.ask("Say hello in 10 languages"); // 2048 tokens, model: GPT_3_5_TURBO_1106
String result2 = gpt.ask("Say hello in 10 languages", 512); // 512 tokens, model: GPT_3_5_TURBO_1106

ㅤ ㅤ ㅤ

Using a proxy:

Proxy proxy = new Proxy(Proxy.Type.HTTP, new InetSocketAddress("ip", port));
ChatGpt gpt = new ChatGpt("api-key", proxy);

ㅤ ㅤ ㅤ

Using a ChatGPT stream (The answer comes in chunks):

ChatGptStream gptStream = new ChatGptStream("api-key"); // You can specify a proxy


gptStream.ask("Hello", new GptStream() {
    @Override
    public void start() {
        System.out.println("Starting");
    }

    @Override
    public void next(int i, String s) { // 'i' is not a chunk index, it is a message index (If you generate several of them at a time)
        System.out.println("Next letters: " + s);
    }

    @Override
    public void end(int i, String s) {
        System.out.println("Result: " + s);
    }

    @Override
    public void error(RuntimeException e) {
        e.printStackTrace();
     }
});

ㅤ ㅤ ㅤ

Request with different parameters (Not everything is shown here):

ChatGpt gpt = new ChatGpt("api-key");

ChatGptRequestDTO request = ChatGptRequestDTO.builder()
        .maxTokens(1024) // Max tokens
        .model(ChatGptModel.GPT_3_5_TURBO.getName()) // Model
        .user("User") // User name (Optional)
        .n(3) // Generation count (Optional)
        .messages(
                List.of(
                        ChatGptMessage.builder()
                                .role(ChatGptRole.USER.getName())
                                .content("Hello, my name is Nick")
                                .build(),
                        ChatGptMessage.builder()
                                .role(ChatGptRole.USER.getName())
                                .content("What is my name?")
                                .build()
        )) // Messages
        .build();

List<String> result = gpt.ask(request);

ㅤ ㅤ ㅤ

Possible exception:

BadRequestException.class // Bad request
InvalidAuthenticationException.class // Failed to authenticate (Perhaps the key is incorrect)
ToManyRequestException.class // Most likely you have exceeded the message limit for a certain time
ServerException.class // Server exception
ServiceUnavailableException.class // Most likely the server is overloaded, try again
CustomHttpException.class  // Ohter exceptions

They all extend from:
        HttpException.class

ㅤ ㅤ ㅤ

If you have several keys, you can use:

ㅤ Single key container:

ApiKeysContainer apiKeysContainer = ApiKeysContainer.of("api-key");
        
ChatGpt gpt = new ChatGpt(apiKeysContainer);

ㅤ Random select container:

RandomSelectApiKeysContainer apiKeysContainer = new RandomSelectApiKeysContainer(List.of("api-key-1", "api-key-2"));

ChatGpt gpt = new ChatGpt(apiKeysContainer);

ㅤ Rounded select container:

RoundSelectApiKeyContainer apiKeysContainer = new RoundSelectApiKeyContainer(List.of("api-key-1", "api-key-2"));

ChatGpt gpt = new ChatGpt(apiKeysContainer);