Chat Completions

Note

调用前请先构建客户端，构建代码如下：

OpenAiClient client = OpenAiClient.builder()
        .apiHost("https://api.openai.com")
        .apiKey(System.getProperty("openai.token"))
        .build();

System.getProperty("openai.token") 是访问 API 授权的关键。

Create chat completion

---

为给定的聊天对话创建模型响应。

List<CompletionMessageEntity> messages = Lists.newArrayList();
messages.add(CompletionMessageEntity.builder()
    .content("Hello, my name is openai-java-sdk")
    .build());

CompletionChatEntity configure = CompletionChatEntity.builder()
    .messages(messages)
    .build();

client.createChatCompletion(configure)
    .getChoices()
    .forEach(choice -> messages.add(choice.getMessage()));

messages.add(CompletionMessageEntity.builder()
    .content("What is my name?")
    .build());

client.createChatCompletion(configure)
    .getChoices()
    .stream()
    .map(v -> v.getMessage().getContent())
    .collect(Collectors.toList())
    .toString()
    .contains("openai-java-sdk");

Body:

Name	Type	Required
model	String	Yes
prompt	String	Yes
temperature	Number	No
maxTokens	Number	No
topP	Number	No
bestOf	Number	No
frequencyPenalty	Number	No
presencePenalty	Number	No
stop	List<String>	No

Returns:

{
  "id": "chatcmpl-123",
  "object": "chat.completion",
  "created": 1677652288,
  "choices": [
    {
      "index": 0,
      "message": {
        "role": "assistant",
        "content": "\n\nHello there, how may I assist you today?"
      },
      "finish_reason": "stop"
    }
  ],
  "usage": {
    "prompt_tokens": 9,
    "completion_tokens": 12,
    "total_tokens": 21
  }
}