Chat Model Examples

import dev.langchain4j.model.vertexai.VertexAiChatModel;
import dev.langchain4j.data.message.UserMessage;
import dev.langchain4j.model.chat.request.ChatRequest;
import dev.langchain4j.model.chat.response.ChatResponse;

VertexAiChatModel model = VertexAiChatModel.builder()
    .endpoint("https://us-central1-aiplatform.googleapis.com/v1/")
    .project("your-project-id")
    .location("us-central1")
    .publisher("google")
    .modelName("chat-bison@001")
    .build();

ChatRequest request = ChatRequest.builder()
    .messages(UserMessage.from("What is artificial intelligence?"))
    .build();

ChatResponse response = model.chat(request);
System.out.println(response.aiMessage().text());

VertexAiChatModel model = VertexAiChatModel.builder()
    .endpoint("https://us-central1-aiplatform.googleapis.com/v1/")
    .project("your-project-id")
    .location("us-central1")
    .publisher("google")
    .modelName("chat-bison@001")
    .temperature(0.7)              // Controls randomness (0.0-1.0)
    .maxOutputTokens(500)          // Maximum response length
    .topK(40)                      // Top-K sampling
    .topP(0.95)                    // Nucleus sampling
    .maxRetries(3)                 // Retry attempts on failure
    .build();

import dev.langchain4j.data.message.ChatMessage;
import dev.langchain4j.data.message.UserMessage;
import dev.langchain4j.data.message.AiMessage;
import java.util.ArrayList;
import java.util.List;

List<ChatMessage> messages = new ArrayList<>();
messages.add(UserMessage.from("What is Java?"));
messages.add(AiMessage.from("Java is a high-level, object-oriented programming language."));
messages.add(UserMessage.from("What are its main features?"));

ChatRequest request = ChatRequest.builder()
    .messages(messages)
    .build();

ChatResponse response = model.chat(request);
System.out.println(response.aiMessage().text());

import java.util.Scanner;

List<ChatMessage> conversation = new ArrayList<>();
Scanner scanner = new Scanner(System.in);

while (true) {
    System.out.print("You: ");
    String userInput = scanner.nextLine();

    if (userInput.equalsIgnoreCase("exit")) {
        break;
    }

    conversation.add(UserMessage.from(userInput));

    ChatRequest request = ChatRequest.builder()
        .messages(conversation)
        .build();

    ChatResponse response = model.chat(request);
    String aiResponse = response.aiMessage().text();

    conversation.add(AiMessage.from(aiResponse));
    System.out.println("AI: " + aiResponse);
}

import com.google.auth.oauth2.GoogleCredentials;
import java.io.FileInputStream;

GoogleCredentials credentials = GoogleCredentials.fromStream(
    new FileInputStream("/path/to/service-account.json")
);

VertexAiChatModel model = VertexAiChatModel.builder()
    .endpoint("https://us-central1-aiplatform.googleapis.com/v1/")
    .project("your-project-id")
    .location("us-central1")
    .publisher("google")
    .modelName("chat-bison@001")
    .credentials(credentials)
    .build();

VertexAiChatModel deterministicModel = VertexAiChatModel.builder()
    .endpoint("https://us-central1-aiplatform.googleapis.com/v1/")
    .project("your-project-id")
    .location("us-central1")
    .publisher("google")
    .modelName("chat-bison@001")
    .temperature(0.0)  // Most deterministic, focused
    .build();

VertexAiChatModel creativeModel = VertexAiChatModel.builder()
    .endpoint("https://us-central1-aiplatform.googleapis.com/v1/")
    .project("your-project-id")
    .location("us-central1")
    .publisher("google")
    .modelName("chat-bison@001")
    .temperature(1.0)  // Most creative, varied
    .build();

VertexAiChatModel balancedModel = VertexAiChatModel.builder()
    .endpoint("https://us-central1-aiplatform.googleapis.com/v1/")
    .project("your-project-id")
    .location("us-central1")
    .publisher("google")
    .modelName("chat-bison@001")
    .temperature(0.7)  // Balanced creativity and focus
    .build();

VertexAiChatModel model = VertexAiChatModel.builder()
    .endpoint("https://us-central1-aiplatform.googleapis.com/v1/")
    .project("your-project-id")
    .location("us-central1")
    .publisher("google")
    .modelName("chat-bison@001")
    .maxOutputTokens(100)  // Limit response to 100 tokens
    .build();

ChatRequest request = ChatRequest.builder()
    .messages(UserMessage.from("Write a long essay about AI"))
    .build();

ChatResponse response = model.chat(request);
// Response will be truncated to ~100 tokens

// Focused responses with Top-K
VertexAiChatModel focusedModel = VertexAiChatModel.builder()
    .endpoint("https://us-central1-aiplatform.googleapis.com/v1/")
    .project("your-project-id")
    .location("us-central1")
    .publisher("google")
    .modelName("chat-bison@001")
    .topK(10)    // Consider only top 10 tokens
    .topP(0.9)   // Nucleus sampling threshold
    .build();

VertexAiChatModel geminiModel = VertexAiChatModel.builder()
    .endpoint("https://us-central1-aiplatform.googleapis.com/v1/")
    .project("your-project-id")
    .location("us-central1")
    .publisher("google")
    .modelName("gemini-pro")
    .build();

VertexAiChatModel latestModel = VertexAiChatModel.builder()
    .endpoint("https://us-central1-aiplatform.googleapis.com/v1/")
    .project("your-project-id")
    .location("us-central1")
    .publisher("google")
    .modelName("chat-bison@002")  // Latest version
    .build();

try {
    ChatResponse response = model.chat(request);
    System.out.println(response.aiMessage().text());
} catch (Exception e) {
    System.err.println("Chat failed: " + e.getMessage());
    // Model automatically retries transient errors up to maxRetries
    // This catches non-retryable errors only
}

ChatResponse response = model.chat(request);

String text = response.aiMessage().text();
TokenUsage tokenUsage = response.tokenUsage();
FinishReason finishReason = response.finishReason();

System.out.println("Response: " + text);
System.out.println("Input tokens: " + tokenUsage.inputTokenCount());
System.out.println("Output tokens: " + tokenUsage.outputTokenCount());
System.out.println("Total tokens: " + tokenUsage.totalTokenCount());
System.out.println("Finish reason: " + finishReason);