Spring AI Chat

from("direct:chat")
    .to("spring-ai-chat:test?chatModel=#chatModel");

// Usage:
String response = template.requestBody("direct:chat", "What is the capital of Italy?", String.class);

from("direct:chat")
    .to("spring-ai-chat:test?chatModel=#chatModel&chatOperation=CHAT_SINGLE_MESSAGE_WITH_PROMPT");

// Usage:
String promptTemplate = "Create a recipe for {{dish}} using {{ingredients}}";
Map<String, Object> variables = Map.of(
    "dish", "pasta carbonara",
    "ingredients", "eggs, cheek lard, parmesan, black pepper"
);

String response = template.requestBodyAndHeader(
    "direct:chat",
    variables,
    SpringAiChatConstants.PROMPT_TEMPLATE,
    promptTemplate,
    String.class
);

import org.springframework.ai.chat.messages.*;

from("direct:chat")
    .to("spring-ai-chat:test?chatModel=#chatModel&chatOperation=CHAT_MULTIPLE_MESSAGES");

// Usage:
List<Message> conversation = new ArrayList<>();
conversation.add(new SystemMessage("You are a helpful cooking assistant."));
conversation.add(new UserMessage("What's a good pasta recipe?"));
conversation.add(new AssistantMessage("I recommend pasta carbonara..."));
conversation.add(new UserMessage("Can I make it vegetarian?"));

String response = template.requestBody("direct:chat", conversation, String.class);

import org.springframework.ai.vectorstore.VectorStore;
import org.springframework.ai.vectorstore.SimpleVectorStore;

// Create and populate vector store
VectorStore vectorStore = new SimpleVectorStore();
// ... add documents to vector store ...

context.getRegistry().bind("vectorStore", vectorStore);

from("direct:chat")
    .to("spring-ai-chat:test?chatModel=#chatModel&vectorStore=#vectorStore&topK=5");

import org.apache.camel.component.springai.chat.rag.SpringAiRagAggregatorStrategy;

SpringAiRagAggregatorStrategy aggregatorStrategy = new SpringAiRagAggregatorStrategy();

from("direct:chat")
    .enrich("direct:retrieve-context", aggregatorStrategy)
    .to("spring-ai-chat:test?chatModel=#chatModel");

from("direct:retrieve-context")
    .setBody(simple("Retrieved context based on question ${body}"));

import org.springframework.ai.document.Document;
import static org.apache.camel.component.springai.chat.SpringAiChatConstants.AUGMENTED_DATA;

List<Document> context = List.of(
    new Document("The Eiffel Tower is located in Paris, France."),
    new Document("It was built in 1889 for the World's Fair.")
);

String response = template.requestBodyAndHeader(
    "direct:chat",
    "Where is the Eiffel Tower located?",
    AUGMENTED_DATA,
    context,
    String.class
);

from("direct:chat")
    .to("spring-ai-chat:test?chatModel=#chatModel&temperature=0.8&maxTokens=500");

// Using URL parameters - nested property syntax
from("direct:chat")
    .to("spring-ai-chat:test?chatModel=#chatModel"
        + "&userMetadata.messageId=msg-123"
        + "&userMetadata.userId=user-456"
        + "&userMetadata.priority=high");

// Using URL parameters - nested property syntax
from("direct:chat")
    .to("spring-ai-chat:test?chatModel=#chatModel"
        + "&systemMetadata.version=1.0"
        + "&systemMetadata.model=gpt-4"
        + "&systemMetadata.promptVersion=2024-01");

// Create chat memory with message window
ChatMemory chatMemory = MessageWindowChatMemory.builder()
    .chatMemoryRepository(new InMemoryChatMemoryRepository())
    .maxMessages(10)  // Keep last 10 messages
    .build();

context.getRegistry().bind("chatMemory", chatMemory);

// Configure endpoint with memory
from("direct:chat")
    .to("spring-ai-chat:memory?chatModel=#chatModel&chatMemory=#chatMemory");

// Conversation ID for isolation
template.requestBodyAndHeader("direct:chat", "My name is Alice",
    SpringAiChatConstants.CONVERSATION_ID, "user-123", String.class);

template.requestBodyAndHeader("direct:chat", "What's my name?",
    SpringAiChatConstants.CONVERSATION_ID, "user-123", String.class); // Will remember: "Alice"

OllamaApi ollamaApi = OllamaApi.builder()
                .baseUrl(OLLAMA.baseUrl())
                .build();

EmbeddingModel embeddingModel = OllamaEmbeddingModel.builder()
        .ollamaApi(ollamaApi)
        .defaultOptions(OllamaEmbeddingOptions.builder()
                .model("embeddinggemma:300m")
                .build())
        .build();

// Create vector store for chat memory
chatMemoryVectorStore = SimpleVectorStore.builder(embeddingModel)
        .build();

// Create vector store for chat memory
VectorStore chatMemoryVectorStore = SimpleVectorStore.builder(embeddingModel)
    .build();

context.getRegistry().bind("chatMemoryVectorStore", chatMemoryVectorStore);

// Configure endpoint with vector memory
from("direct:vector-memory")
    .to("spring-ai-chat:vmem?chatModel=#chatModel&chatMemoryVectorStore=#chatMemoryVectorStore");

// Memory is managed automatically with semantic search
template.requestBodyAndHeader("direct:vector-memory",
    "My favorite programming language is Java",
    SpringAiChatConstants.CONVERSATION_ID, "user-123", String.class);

template.requestBodyAndHeader("direct:vector-memory",
    "What programming language did I mention?",
    SpringAiChatConstants.CONVERSATION_ID, "user-123", String.class); // Will use semantic search to retrieve: "Java"

ChatClient chatClient = ChatClient.builder(chatModel)
    .defaultAdvisors(new MessageChatMemoryAdvisor(chatMemory))
    .defaultSystem("You are a helpful assistant")
    .build();

from("direct:chat")
    .to("spring-ai-chat:client?chatClient=#chatClient");