giuseppe-trisciuoglio/developer-kit

aws-sdk-java-v2-bedrock

Provides Amazon Bedrock patterns using AWS SDK for Java 2.x.

소스 보기

사용 방법

  1. 설치 명령을 복사하고 자신의 에이전트 환경에 설치하기 전에 소스 패키지를 확인하세요.
  2. 필요한 도구, 인증 정보, 지원 환경은 원본 문서에서 확인하세요.
  3. 입력과 원하는 산출물을 설명하고 출처와 결과를 대조하세요.
npx skills add https://github.com/giuseppe-trisciuoglio/developer-kit --skill aws-sdk-java-v2-bedrock

시작 전 확인

  • 설치에는 모델 구독, API 한도, 파일 접근 권한이 포함되지 않습니다.
  • 별점 수와 설치 수는 활동 지표이며 독립적인 품질·보안 검증이 아닙니다.
원본 Skill 문서

원본 저장소의 제목, 예시, 코드, 표, 링크, 이미지를 유지해 표시합니다.

AWS SDK for Java 2.x - Amazon Bedrock

Overview

Invokes foundation models through AWS SDK for Java 2.x. Configures clients, builds model-specific JSON payloads, handles streaming responses with error recovery, creates embeddings for RAG, integrates generative AI into Spring Boot applications, and implements exponential backoff for resilience.

When to Use

  • Invoke Claude, Llama, Titan, or Stable Diffusion for text/image generation
  • Configure BedrockClient and BedrockRuntimeClient instances
  • Build and parse model-specific payloads (Claude, Titan, Llama formats)
  • Stream real-time AI responses with async handlers and error recovery
  • Create embeddings for retrieval-augmented generation
  • Integrate generative AI into Spring Boot microservices
  • Handle throttling with exponential backoff retry logic

Quick Start

Dependencies

xml
<!-- Bedrock (model management) -->
<dependency>
    <groupId>software.amazon.awssdk</groupId>
    <artifactId>bedrock</artifactId>
</dependency>

<!-- Bedrock Runtime (model invocation) -->
<dependency>
    <groupId>software.amazon.awssdk</groupId>
    <artifactId>bedrockruntime</artifactId>
</dependency>

<!-- For JSON processing -->
<dependency>
    <groupId>org.json</groupId>
    <artifactId>json</artifactId>
    <version>20231013</version>
</dependency>

Client Setup

java
import software.amazon.awssdk.regions.Region;
import software.amazon.awssdk.services.bedrock.BedrockClient;
import software.amazon.awssdk.services.bedrockruntime.BedrockRuntimeClient;

// Model management client
BedrockClient bedrockClient = BedrockClient.builder()
    .region(Region.US_EAST_1)
    .build();

// Model invocation client
BedrockRuntimeClient bedrockRuntimeClient = BedrockRuntimeClient.builder()
    .region(Region.US_EAST_1)
    .build();

Instructions

Follow these steps for production-ready Bedrock integration:

  1. Configure AWS Credentials - Set up IAM roles with Bedrock permissions (avoid access keys)
  2. Enable Model Access - Request access to specific foundation models in AWS Console
  3. Initialize Clients - Create reusable BedrockClient and BedrockRuntimeClient instances
  4. Validate Model Availability - Test with a simple invocation before production use
  5. Build Payloads - Create model-specific JSON payloads with proper format
  6. Handle Responses - Parse response structure and extract content
  7. Implement Streaming - Use response stream handlers for real-time generation
  8. Add Error Handling - Implement retry logic with exponential backoff

Validation Checkpoint: Always test with a simple prompt (e.g., "Hello") before production use to verify model access and response parsing.

Examples

Text Generation with Claude

java
public String generateWithClaude(BedrockRuntimeClient client, String prompt) {
    JSONObject payload = new JSONObject()
        .put("anthropic_version", "bedrock-2023-05-31")
        .put("max_tokens", 1000)
        .put("messages", new JSONObject[]{
            new JSONObject().put("role", "user").put("content", prompt)
        });

    InvokeModelResponse response = client.invokeModel(InvokeModelRequest.builder()
        .modelId("anthropic.claude-sonnet-4-5-20250929-v1:0")
        .body(SdkBytes.fromUtf8String(payload.toString()))
        .build());

    JSONObject responseBody = new JSONObject(response.body().asUtf8String());
    return responseBody.getJSONArray("content")
        .getJSONObject(0)
        .getString("text");
}

Model Discovery

java
import software.amazon.awssdk.services.bedrock.model.*;

public List<FoundationModelSummary> listFoundationModels(BedrockClient bedrockClient) {
    return bedrockClient.listFoundationModels().modelSummaries();
}

Multi-Model Invocation

java
public String invokeModel(BedrockRuntimeClient client, String modelId, String prompt) {
    JSONObject payload = createPayload(modelId, prompt);

    InvokeModelResponse response = client.invokeModel(request -> request
        .modelId(modelId)
        .body(SdkBytes.fromUtf8String(payload.toString())));

    return extractTextFromResponse(modelId, response.body().asUtf8String());
}

private JSONObject createPayload(String modelId, String prompt) {
    if (modelId.startsWith("anthropic.claude")) {
        return new JSONObject()
            .put("anthropic_version", "bedrock-2023-05-31")
            .put("max_tokens", 1000)
            .put("messages", new JSONObject[]{
                new JSONObject().put("role", "user").put("content", prompt)
            });
    } else if (modelId.startsWith("amazon.titan")) {
        return new JSONObject()
            .put("inputText", prompt)
            .put("textGenerationConfig", new JSONObject()
                .put("maxTokenCount", 512)
                .put("temperature", 0.7));
    } else if (modelId.startsWith("meta.llama")) {
        return new JSONObject()
            .put("prompt", "[INST] " + prompt + " [/INST]")
            .put("max_gen_len", 512)
            .put("temperature", 0.7);
    }
    throw new IllegalArgumentException("Unsupported model: " + modelId);
}

Streaming Response with Error Handling

java
public String streamResponseWithRetry(BedrockRuntimeClient client, String modelId, String prompt, int maxRetries) {
    int attempt = 0;
    while (attempt < maxRetries) {
        try {
            JSONObject payload = createPayload(modelId, prompt);
            StringBuilder fullResponse = new StringBuilder();

            InvokeModelWithResponseStreamRequest request = InvokeModelWithResponseStreamRequest.builder()
                .modelId(modelId)
                .body(SdkBytes.fromUtf8String(payload.toString()))
                .build();

            client.invokeModelWithResponseStream(request,
                InvokeModelWithResponseStreamResponseHandler.builder()
                    .onEventStream(stream -> stream.forEach(event -> {
                        if (event instanceof PayloadPart) {
                            String chunk = ((PayloadPart) event).bytes().asUtf8String();
                            fullResponse.append(chunk);
                        }
                    }))
                    .onError(e -> System.err.println("Stream error: " + e.getMessage()))
                    .build());

            return fullResponse.toString();
        } catch (Exception e) {
            attempt++;
            if (attempt >= maxRetries) {
                throw new RuntimeException("Stream failed after " + maxRetries + " attempts", e);
            }
            try {
                Thread.sleep((long) Math.pow(2, attempt) * 1000); // Exponential backoff
            } catch (InterruptedException ie) {
                Thread.currentThread().interrupt();
                throw new RuntimeException("Interrupted during retry", ie);
            }
        }
    }
    throw new RuntimeException("Unexpected error in streaming");
}

Exponential Backoff for Throttling

java
import software.amazon.awssdk.awscore.exception.AwsServiceException;

public <T> T invokeWithRetry(Supplier<T> invocation, int maxRetries) {
    int attempt = 0;
    while (attempt < maxRetries) {
        try {
            return invocation.get();
        } catch (AwsServiceException e) {
            if (e.statusCode() == 429 || e.statusCode() >= 500) {
                attempt++;
                if (attempt >= maxRetries) throw e;
                long delayMs = Math.min(1000 * (1L << attempt) + (long) (Math.random() * 1000), 30000);
                Thread.sleep(delayMs);
            } else {
                throw e;
            }
        }
    }
    throw new IllegalStateException("Should not reach here");
}

Text Embeddings

java
public double[] createEmbeddings(BedrockRuntimeClient client, String text) {
    String modelId = "amazon.titan-embed-text-v1";

    JSONObject payload = new JSONObject().put("inputText", text);

    InvokeModelResponse response = client.invokeModel(request -> request
        .modelId(modelId)
        .body(SdkBytes.fromUtf8String(payload.toString())));

    JSONObject responseBody = new JSONObject(response.body().asUtf8String());
    JSONArray embeddingArray = responseBody.getJSONArray("embedding");

    double[] embeddings = new double[embeddingArray.length()];
    for (int i = 0; i < embeddingArray.length(); i++) {
        embeddings[i] = embeddingArray.getDouble(i);
    }
    return embeddings;
}

Spring Boot Integration

java
@Configuration
public class BedrockConfiguration {

    @Bean
    public BedrockClient bedrockClient() {
        return BedrockClient.builder()
            .region(Region.US_EAST_1)
            .build();
    }

    @Bean
    public BedrockRuntimeClient bedrockRuntimeClient() {
        return BedrockRuntimeClient.builder()
            .region(Region.US_EAST_1)
            .build();
    }
}

@Service
public class BedrockAIService {

    private final BedrockRuntimeClient bedrockRuntimeClient;
    private final ObjectMapper mapper;

    @Value("${bedrock.default-model-id:anthropic.claude-sonnet-4-5-20250929-v1:0}")
    private String defaultModelId;

    public BedrockAIService(BedrockRuntimeClient bedrockRuntimeClient, ObjectMapper mapper) {
        this.bedrockRuntimeClient = bedrockRuntimeClient;
        this.mapper = mapper;
    }

    public String generateText(String prompt) {
        Map<String, Object> payload = Map.of(
            "anthropic_version", "bedrock-2023-05-31",
            "max_tokens", 1000,
            "messages", List.of(Map.of("role", "user", "content", prompt))
        );

        InvokeModelResponse response = bedrockRuntimeClient.invokeModel(
            InvokeModelRequest.builder()
                .modelId(defaultModelId)
                .body(SdkBytes.fromUtf8String(mapper.writeValueAsString(payload)))
                .build());

        return extractText(response.body().asUtf8String());
    }
}

See examples directory for comprehensive usage patterns.

Best Practices

Model Selection

  • Claude 4.5 Sonnet: Complex reasoning, analysis, and creative tasks
  • Claude 4.5 Haiku: Fast and affordable for real-time applications
  • Llama 3.1: Open-source alternative for general tasks
  • Titan: AWS native, cost-effective for simple text generation

Performance

  • Reuse client instances (avoid creating new clients per request)
  • Use async clients for I/O operations
  • Implement streaming for long responses
  • Cache foundation model lists

Security

  • Never log sensitive prompt data
  • Use IAM roles for authentication
  • Sanitize user inputs to prevent prompt injection
  • Implement rate limiting for public applications

Constraints and Warnings

  • Cost Management: Bedrock API calls incur charges per token; implement usage monitoring and budget alerts.
  • Model Access: Foundation models must be enabled in AWS Console; verify region availability.
  • Rate Limits: Implement exponential backoff for throttling; check per-model limits.
  • Payload Size: Maximum payload size varies by model; use chunking for large documents.
  • Streaming Complexity: Handle partial content and error recovery carefully.
  • Data Privacy: Prompts and responses may be logged by AWS; review data policies.
  • Credentials: Never embed credentials in code; use IAM roles for EC2/Lambda.

Common Model IDs

  • Claude Sonnet 4.5: anthropic.claude-sonnet-4-5-20250929-v1:0
  • Claude Haiku 4.5: anthropic.claude-haiku-4-5-20251001-v1:0
  • Llama 3.1 70B: meta.llama3-1-70b-instruct-v1:0
  • Titan Embeddings: amazon.titan-embed-text-v1

See Model Reference for complete list.

References

Related Skills

  • aws-sdk-java-v2-core - Core AWS SDK patterns
  • langchain4j-ai-services-patterns - LangChain4j integration
  • spring-boot-dependency-injection - Spring DI patterns

관련 작업

같은 저장소의 Skills

더 많은 Skills

모든 Skills
giuseppe-trisciuoglio
커뮤니티

drizzle-orm-patterns

Provides comprehensive Drizzle ORM patterns for schema definition, CRUD operations, relations, queries, transactions, and migrations. Proactively use for any Drizzle ORM development including defining database schemas, writing type-safe queries, implementing relations, managing transactions, and setting up migrations with Drizzle Kit. Supports PostgreSQL, MySQL, SQLite, MSSQL, and CockroachDB.

1시간 설치 수
3
GitHub Stars
372
업데이트
9월 10일
giuseppe-trisciuoglio
커뮤니티

tailwind-css-patterns

Provides comprehensive Tailwind CSS utility-first styling patterns including responsive design, layout utilities, flexbox, grid, spacing, typography, colors, and modern CSS best practices. Use when styling React/Vue/Svelte components, building responsive layouts, implementing design systems, or optimizing CSS workflow.

1시간 설치 수
1
GitHub Stars
372
업데이트
9월 10일
giuseppe-trisciuoglio
커뮤니티

dynamodb-toolbox-patterns

Provides TypeScript patterns for DynamoDB-Toolbox v2 including schema/table/entity modeling, .build() command workflow, query/scan access patterns, batch and transaction operations, and single-table design with computed keys. Use when implementing type-safe DynamoDB access layers with DynamoDB-Toolbox v2 in TypeScript services or serverless applications.

1시간 설치 수
1
GitHub Stars
371
업데이트
9월 10일
giuseppe-trisciuoglio
커뮤니티

spring-boot-dependency-injection

Provides dependency injection patterns for Spring Boot projects, including constructor-first design, optional collaborator handling, bean selection, and wiring validation. Use when creating services and configurations, replacing field injection, or troubleshooting ambiguous or fragile Spring wiring.

1시간 설치 수
1
GitHub Stars
371
업데이트
9월 10일