本文為英文版的機器翻譯版本，如內容有任何歧義或不一致之處，概以英文版為準。

# 使用 AWS SDKs 的 Amazon Bedrock 執行期案例
<a name="service_code_examples_bedrock-runtime_scenarios"></a>

下列程式碼範例示範如何在 Amazon Bedrock Runtime AWS SDKs中實作常見案例。這些案例向您示範如何呼叫 Amazon Bedrock 執行時期中的多個函數或與其他 AWS 服務組合，以完成特定任務。每個案例均包含完整原始碼的連結，您可在連結中找到如何設定和執行程式碼的相關指示。

案例的目標是獲得中等水平的經驗，協助您了解內容中的服務動作。

**Topics**
+ [建立演練場應用程式，與 Amazon Bedrock 基礎模型互動](bedrock-runtime_example_cross_FMPlayground_section.md)
+ [建立和調用受管提示](bedrock-runtime_example_bedrock-agent_GettingStartedWithBedrockPrompts_section.md)
+ [使用 Amazon Bedrock 從文字提示產生影片](bedrock-runtime_example_bedrock-runtime_Scenario_GenerateVideos_NovaReel_section.md)
+ [在 Amazon Bedrock 上調用多個基礎模型](bedrock-runtime_example_bedrock-runtime_Scenario_InvokeModels_section.md)
+ [使用 Step Functions 協調生成式 AI 應用程式](bedrock-runtime_example_cross_ServerlessPromptChaining_section.md)
+ [搭配 Converse API 使用工具](bedrock-runtime_example_bedrock-runtime_Scenario_ToolUse_section.md)

# 建立提供遊樂場的範例應用程式，以使用 AWS SDK 與 Amazon Bedrock 基礎模型互動
<a name="bedrock-runtime_example_cross_FMPlayground_section"></a>

下列程式碼範例示範如何建立演練場，透過不同的模式與 Amazon Bedrock 基礎模型互動。

------
#### [ .NET ]

**適用於 .NET 的 SDK**  
 .NET 基礎模型 (FM) 演練場是 .NET MAUI Blazor 範例應用程式，展示如何從 C\$1 程式碼使用 Amazon Bedrock。此範例顯示 .NET 和 C\$1 開發人員如何使用 Amazon Bedrock 建置啟用生成式 AI 技術的應用程式。您可以使用下列四個演練場，測試 Amazon Bedrock 基礎模型並與之互動：  
+ 文字演練場。
+ 聊天演練場。
+ 語音聊天演練場。
+ 影像演練場。
此範例也列出並顯示您有權存取的基礎模型及其特性。如需原始程式碼和部署的說明，請參閱 [GitHub](https://github.com/build-on-aws/dotnet-fm-playground) 中的專案。  

**此範例中使用的服務**
+ Amazon Bedrock 執行時期

------
#### [ Java ]

**適用於 Java 2.x 的 SDK**  
 Java 基礎模型 (FM) 演練場是 Spring Boot 範例應用程式，展示如何搭配 Java 使用 Amazon Bedrock。此範例顯示 Java 開發人員如何使用 Amazon Bedrock 建置啟用生成式 AI 技術的應用程式。您可以使用下列三個演練場，測試 Amazon Bedrock 基礎模型並與之互動：  
+ 文字演練場。
+ 聊天演練場。
+ 影像演練場。
此範例也列出並顯示您有權存取的基礎模型及其特性。如需原始程式碼和部署的說明，請參閱 [GitHub](https://github.com/build-on-aws/java-fm-playground) 中的專案。  

**此範例中使用的服務**
+ Amazon Bedrock 執行時期

------
#### [ Python ]

**適用於 Python 的 SDK (Boto3)**  
 Python 基礎模型 (FM) 演練場是 Python/FastAPI 範例應用程式，展示如何搭配 Python 使用 Amazon Bedrock。此範例示範 Python 開發人員如何使用 Amazon Bedrock 建置啟用生成式 AI 技術的應用程式。您可以使用下列三個演練場，測試 Amazon Bedrock 基礎模型並與之互動：  
+ 文字演練場。
+ 聊天演練場。
+ 影像演練場。
此範例也列出並顯示您有權存取的基礎模型及其特性。如需原始程式碼和部署的說明，請參閱 [GitHub](https://github.com/build-on-aws/python-fm-playground) 中的專案。  

**此範例中使用的服務**
+ Amazon Bedrock 執行時期

------

如需 AWS SDK 開發人員指南和程式碼範例的完整清單，請參閱 [搭配 AWS SDK 使用 Amazon Bedrock](sdk-general-information-section.md)。此主題也包含有關入門的資訊和舊版 SDK 的詳細資訊。

# end-to-end範例，示範如何使用 AWS SDK 建立和叫用 Amazon Bedrock 受管提示
<a name="bedrock-runtime_example_bedrock-agent_GettingStartedWithBedrockPrompts_section"></a>

以下程式碼範例顯示做法：
+ 建立受管提示。
+ 建立提示的版本。
+ 使用該版本調用提示。
+ 清除資源 (選用)。

------
#### [ Python ]

**適用於 Python 的 SDK (Boto3)**  
 GitHub 上提供更多範例。尋找完整範例，並了解如何在 [AWS 程式碼範例儲存庫](https://github.com/awsdocs/aws-doc-sdk-examples/tree/main/python/example_code/bedrock-agent#code-examples)中設定和執行。
建立和調用受管提示。  

```
import argparse
import boto3
import logging
import time

# Now import the modules
from prompt import create_prompt, create_prompt_version, delete_prompt
from run_prompt import invoke_prompt

logging.basicConfig(
    level=logging.INFO,
    format='%(levelname)s: %(message)s'
)
logger = logging.getLogger(__name__)



def run_scenario(bedrock_client, bedrock_runtime_client, model_id, cleanup=True):
    """
    Runs the Amazon Bedrock managed prompt scenario.
    
    Args:
        bedrock_client: The Amazon Bedrock Agent client.
        bedrock_runtime_client: The Amazon Bedrock Runtime client.
        model_id (str): The model ID to use for the prompt.
        cleanup (bool): Whether to clean up resources at the end of the scenario.
        
    Returns:
        dict: A dictionary containing the created resources.
    """
    prompt_id = None
    
    try:
        # Step 1: Create a prompt
        print("\n=== Step 1: Creating a prompt ===")
        prompt_name = f"PlaylistGenerator-{int(time.time())}"
        prompt_description = "Playlist generator"
        prompt_template = """
          Make me a {{genre}} playlist consisting of the following number of songs: {{number}}."""
        
        create_response = create_prompt(
            bedrock_client,
            prompt_name,
            prompt_description,
            prompt_template,
            model_id
        )
        
        prompt_id = create_response['id']
        print(f"Created prompt: {prompt_name} with ID: {prompt_id}")
        
        # Create a version of the prompt
        print("\n=== Creating a version of the prompt ===")
        version_response = create_prompt_version(
            bedrock_client,
            prompt_id,
            description="Initial version of the product description generator"
        )
        
        prompt_version_arn = version_response['arn']
        prompt_version = version_response['version']

        print(f"Created prompt version: {prompt_version}")
        print(f"Prompt version ARN: {prompt_version_arn}")
        
        # Step 2: Invoke the prompt directly
        print("\n=== Step 2: Invoking the prompt ===")
        input_variables = {
            "genre": "pop",
            "number": "2",
           }
        
        # Use the ARN from the create_prompt_version response
        result = invoke_prompt(
            bedrock_runtime_client,
            prompt_version_arn,  
            input_variables
        )
        # Display the playlist
        print(f"\n{result}")
    
        
        # Step 3: Clean up resources (optional)
        if cleanup:
            print("\n=== Step 3: Cleaning up resources ===")
            
            # Delete the prompt
            print(f"Deleting prompt {prompt_id}...")
            delete_prompt(bedrock_client, prompt_id)
            
            print("Cleanup complete")
        else:
            print("\n=== Resources were not cleaned up ===")
            print(f"Prompt ID: {prompt_id}")
        
   
        
    except Exception as e:
        logger.exception("Error in scenario: %s", str(e))
        
        # Attempt to clean up if an error occurred and cleanup was requested
        if cleanup and prompt_id:
            try:
                print("\nCleaning up resources after error...")
                
                # Delete the prompt
                try:
                    delete_prompt(bedrock_client, prompt_id)
                    print("Cleanup after error complete")
                except Exception as cleanup_error:
                    logger.error("Error during cleanup: %s", str(cleanup_error))
            except Exception as final_error:
                logger.error("Final error during cleanup: %s", str(final_error))
        
        # Re-raise the original exception
        raise

def main():
    """
    Entry point for the Amazon Bedrock managed prompt scenario.
    """
    parser = argparse.ArgumentParser(
        description="Run the Amazon Bedrock managed prompt scenario."
    )
    parser.add_argument(
        '--region',
        default='us-east-1',
        help="The AWS Region to use."
    )
    parser.add_argument(
        '--model-id',
        default='anthropic.claude-v2',
        help="The model ID to use for the prompt."
    )
    parser.add_argument(
        '--cleanup',
        action='store_true',
        default=True,
        help="Clean up resources at the end of the scenario."
    )
    parser.add_argument(
        '--no-cleanup',
        action='store_false',
        dest='cleanup',
        help="Don't clean up resources at the end of the scenario."
    )
    args = parser.parse_args()

    bedrock_client = boto3.client('bedrock-agent', region_name=args.region)
    bedrock_runtime_client = boto3.client('bedrock-runtime', region_name=args.region)
    
    print("=== Amazon Bedrock Managed Prompt Scenario ===")
    print(f"Region: {args.region}")
    print(f"Model ID: {args.model_id}")
    print(f"Cleanup resources: {args.cleanup}")
    
    try:
        run_scenario(
            bedrock_client,
            bedrock_runtime_client,
            args.model_id,
            args.cleanup
        )
        
    except Exception as e:
        logger.exception("Error running scenario: %s", str(e))
        
if __name__ == "__main__":
    main()
```
+ 如需 API 詳細資訊，請參閱《適用於 Python (Boto3) 的AWS SDK API 參考》**中的下列主題。
  + [Converse](https://docs.aws.amazon.com/goto/boto3/bedrock-runtime-2023-09-30/Converse)
  + [CreatePrompt](https://docs.aws.amazon.com/goto/boto3/bedrock-agent-2023-12-12/CreatePrompt)
  + [CreatePromptVersion](https://docs.aws.amazon.com/goto/boto3/bedrock-agent-2023-12-12/CreatePromptVersion)
  + [DeletePrompt](https://docs.aws.amazon.com/goto/boto3/bedrock-agent-2023-12-12/DeletePrompt)

------

如需 AWS SDK 開發人員指南和程式碼範例的完整清單，請參閱 [搭配 AWS SDK 使用 Amazon Bedrock](sdk-general-information-section.md)。此主題也包含有關入門的資訊和舊版 SDK 的詳細資訊。

# 使用 Amazon Bedrock 和 Nova-Reel 從文字提示產生影片
<a name="bedrock-runtime_example_bedrock-runtime_Scenario_GenerateVideos_NovaReel_section"></a>

下列程式碼範例示範如何使用 Amazon Bedrock 和 Nova-Reel 模型，從文字提示產生影片的 Spring Boot 應用程式。

------
#### [ Java ]

**適用於 Java 2.x 的 SDK **  
 GitHub 上提供更多範例。尋找完整範例，並了解如何在 [AWS 程式碼範例儲存庫](https://github.com/awsdocs/aws-doc-sdk-examples/tree/main/javav2/example_code/bedrock-runtime#code-examples)中設定和執行。
使用 Amazon Bedrock 和 Nova-Reel 從文字提示產生影片。  

```
import org.springframework.stereotype.Service;
import software.amazon.awssdk.auth.credentials.ProfileCredentialsProvider;
import software.amazon.awssdk.core.document.Document;
import software.amazon.awssdk.regions.Region;
import software.amazon.awssdk.services.bedrockruntime.BedrockRuntimeAsyncClient;
import software.amazon.awssdk.services.bedrockruntime.model.*;

import java.util.concurrent.CompletableFuture;

@Service
public class VideoGenerationService {

    public GenerateVideoResponse generateVideo(String prompt) {

        // add S3 bucket you want to store your generated videos
        String s3Bucket = "s3://mygeneratedvidoenovatest";


        //Create json request as an instance of Document class
        Document novaRequest = prepareDocument(prompt);

        // Create request
        StartAsyncInvokeRequest request = StartAsyncInvokeRequest.builder()
                .modelId("amazon.nova-reel-v1:0")
                .modelInput(novaRequest)
                .outputDataConfig(AsyncInvokeOutputDataConfig.builder()
                        .s3OutputDataConfig(AsyncInvokeS3OutputDataConfig.builder().s3Uri(s3Bucket).build())
                        .build())
                .build();

        try (BedrockRuntimeAsyncClient bedrockClient = getBedrockRuntimeAsyncClient()) {
            CompletableFuture<StartAsyncInvokeResponse> startAsyncInvokeResponseCompletableFuture = bedrockClient.startAsyncInvoke(request);

            //blocking operation to wait for the AWS API response
            StartAsyncInvokeResponse startAsyncInvokeResponse = startAsyncInvokeResponseCompletableFuture.get();
            System.out.println("invocation ARN: " + startAsyncInvokeResponse.invocationArn());

            GenerateVideoResponse response = new GenerateVideoResponse();
            response.setStatus("inProgress");
            response.setExecutionArn(startAsyncInvokeResponse.invocationArn());

            return response;
        } catch (Exception e) {
            System.out.println(e);
            throw new RuntimeException(e);
        }

    }

    public GenerateVideoResponse checkGenerationStatus(String invocationArn) {
        GenerateVideoResponse response = new GenerateVideoResponse();

        try (BedrockRuntimeAsyncClient bedrockClient = getBedrockRuntimeAsyncClient()) {
            //creating async request to fetch status by invocation Arn
            GetAsyncInvokeRequest asyncRequest = GetAsyncInvokeRequest.builder().invocationArn(invocationArn).build();

            CompletableFuture<GetAsyncInvokeResponse> asyncInvoke = bedrockClient.getAsyncInvoke(asyncRequest);

            //blocking operation to wait for the AWS API response
            GetAsyncInvokeResponse asyncInvokeResponse = asyncInvoke.get();
            System.out.println("Invocation status =" + asyncInvokeResponse.statusAsString());

            response.setExecutionArn(invocationArn);
            response.setStatus(asyncInvokeResponse.statusAsString());
            return response;
        } catch (Exception e) {
            e.printStackTrace();
            throw new RuntimeException(e);
        }

    }

    private static BedrockRuntimeAsyncClient getBedrockRuntimeAsyncClient() {
        BedrockRuntimeAsyncClient bedrockClient = BedrockRuntimeAsyncClient.builder()
                .region(Region.US_EAST_1)
                .credentialsProvider(ProfileCredentialsProvider.create())
                .build();
        return bedrockClient;
    }

    private static Document prepareDocument(String prompt) {
        Document textToVideoParams = Document.mapBuilder()
                .putString("text", prompt)
                .build();

        Document videoGenerationConfig = Document.mapBuilder()
                .putNumber("durationSeconds", 6)
                .putNumber("fps", 24)
                .putString("dimension", "1280x720")
                .build();

        Document novaRequest = Document.mapBuilder()
                .putString("taskType", "TEXT_VIDEO")
                .putDocument("textToVideoParams", textToVideoParams)
                .putDocument("videoGenerationConfig", videoGenerationConfig)
                .build();
        return novaRequest;
    }
}
```
+ 如需 API 詳細資訊，請參閱《*AWS SDK for Java 2.x API 參考*》中的下列主題。
  + [GetAsyncInvoke](https://docs.aws.amazon.com/goto/SdkForJavaV2/bedrock-runtime-2023-09-30/GetAsyncInvoke)
  + [StartAsyncInvoke](https://docs.aws.amazon.com/goto/SdkForJavaV2/bedrock-runtime-2023-09-30/StartAsyncInvoke)

------

如需 AWS SDK 開發人員指南和程式碼範例的完整清單，請參閱 [搭配 AWS SDK 使用 Amazon Bedrock](sdk-general-information-section.md)。此主題也包含有關入門的資訊和舊版 SDK 的詳細資訊。

# 在 Amazon Bedrock 上調用多個基礎模型
<a name="bedrock-runtime_example_bedrock-runtime_Scenario_InvokeModels_section"></a>

下列程式碼範例示範如何在 Amazon Bedrock 上準備並傳送提示至各種大型語言模型 (LLM)

------
#### [ Go ]

**SDK for Go V2**  
 GitHub 上提供更多範例。尋找完整範例，並了解如何在 [AWS 程式碼範例儲存庫](https://github.com/awsdocs/aws-doc-sdk-examples/tree/main/gov2/bedrock-runtime#code-examples)中設定和執行。
在 Amazon Bedrock 上調用多個基礎模型。  

```
import (
	"context"
	"encoding/base64"
	"fmt"
	"log"
	"math/rand"
	"os"
	"path/filepath"
	"strings"

	"github.com/aws/aws-sdk-go-v2/aws"
	"github.com/aws/aws-sdk-go-v2/service/bedrockruntime"
	"github.com/awsdocs/aws-doc-sdk-examples/gov2/bedrock-runtime/actions"
	"github.com/awsdocs/aws-doc-sdk-examples/gov2/demotools"
)

// InvokeModelsScenario demonstrates how to use the Amazon Bedrock Runtime client
// to invoke various foundation models for text and image generation
//
// 1. Generate text with Anthropic Claude 2
// 2. Generate text with Meta Llama 2 Chat
// 3. Generate text and asynchronously process the response stream with Anthropic Claude 2
// 4. Generate an image with the Amazon Titan image generation model
type InvokeModelsScenario struct {
	sdkConfig             aws.Config
	invokeModelWrapper    actions.InvokeModelWrapper
	responseStreamWrapper actions.InvokeModelWithResponseStreamWrapper
	questioner            demotools.IQuestioner
}

// NewInvokeModelsScenario constructs an InvokeModelsScenario instance from a configuration.
// It uses the specified config to get a Bedrock Runtime client and create wrappers for the
// actions used in the scenario.
func NewInvokeModelsScenario(sdkConfig aws.Config, questioner demotools.IQuestioner) InvokeModelsScenario {
	client := bedrockruntime.NewFromConfig(sdkConfig)
	return InvokeModelsScenario{
		sdkConfig:             sdkConfig,
		invokeModelWrapper:    actions.InvokeModelWrapper{BedrockRuntimeClient: client},
		responseStreamWrapper: actions.InvokeModelWithResponseStreamWrapper{BedrockRuntimeClient: client},
		questioner:            questioner,
	}
}

// Runs the interactive scenario.
func (scenario InvokeModelsScenario) Run(ctx context.Context) {
	defer func() {
		if r := recover(); r != nil {
			log.Printf("Something went wrong with the demo: %v\n", r)
		}
	}()

	log.Println(strings.Repeat("=", 77))
	log.Println("Welcome to the Amazon Bedrock Runtime model invocation demo.")
	log.Println(strings.Repeat("=", 77))

	log.Printf("First, let's invoke a few large-language models using the synchronous client:\n\n")

	text2textPrompt := "In one paragraph, who are you?"

	log.Println(strings.Repeat("-", 77))
	log.Printf("Invoking Claude with prompt: %v\n", text2textPrompt)
	scenario.InvokeClaude(ctx, text2textPrompt)

	log.Println(strings.Repeat("=", 77))
	log.Printf("Now, let's invoke Claude with the asynchronous client and process the response stream:\n\n")

	log.Println(strings.Repeat("-", 77))
	log.Printf("Invoking Claude with prompt: %v\n", text2textPrompt)
	scenario.InvokeWithResponseStream(ctx, text2textPrompt)

	log.Println(strings.Repeat("=", 77))
	log.Printf("Now, let's create an image with the Amazon Titan image generation model:\n\n")

	text2ImagePrompt := "stylized picture of a cute old steampunk robot"
	seed := rand.Int63n(2147483648)

	log.Println(strings.Repeat("-", 77))
	log.Printf("Invoking Amazon Titan with prompt: %v\n", text2ImagePrompt)
	scenario.InvokeTitanImage(ctx, text2ImagePrompt, seed)

	log.Println(strings.Repeat("=", 77))
	log.Println("Thanks for watching!")
	log.Println(strings.Repeat("=", 77))
}

func (scenario InvokeModelsScenario) InvokeClaude(ctx context.Context, prompt string) {
	completion, err := scenario.invokeModelWrapper.InvokeClaude(ctx, prompt)
	if err != nil {
		panic(err)
	}
	log.Printf("\nClaude     : %v\n", strings.TrimSpace(completion))
}

func (scenario InvokeModelsScenario) InvokeWithResponseStream(ctx context.Context, prompt string) {
	log.Println("\nClaude with response stream:")
	_, err := scenario.responseStreamWrapper.InvokeModelWithResponseStream(ctx, prompt)
	if err != nil {
		panic(err)
	}
	log.Println()
}

func (scenario InvokeModelsScenario) InvokeTitanImage(ctx context.Context, prompt string, seed int64) {
	base64ImageData, err := scenario.invokeModelWrapper.InvokeTitanImage(ctx, prompt, seed)
	if err != nil {
		panic(err)
	}
	imagePath := saveImage(base64ImageData, "amazon.titan-image-generator-v2")
	fmt.Printf("The generated image has been saved to %s\n", imagePath)
}
```
+ 如需 API 詳細資訊，請參閱《*適用於 Go 的 AWS SDK API 參考*》中的下列主題。
  + [InvokeModel](https://pkg.go.dev/github.com/aws/aws-sdk-go-v2/service/bedrockruntime#Client.InvokeModel)
  + [InvokeModelWithResponseStream](https://pkg.go.dev/github.com/aws/aws-sdk-go-v2/service/bedrockruntime#Client.InvokeModelWithResponseStream)

------
#### [ JavaScript ]

**適用於 JavaScript (v3) 的 SDK**  
 GitHub 上提供更多範例。尋找完整範例，並了解如何在 [AWS 程式碼範例儲存庫](https://github.com/awsdocs/aws-doc-sdk-examples/tree/main/javascriptv3/example_code/bedrock-runtime#code-examples)中設定和執行。

```
import { fileURLToPath } from "node:url";
import {
  Scenario,
  ScenarioAction,
  ScenarioInput,
  ScenarioOutput,
} from "@aws-doc-sdk-examples/lib/scenario/index.js";
import { FoundationModels } from "../config/foundation_models.js";

/**
 * @typedef {Object} ModelConfig
 * @property {Function} module
 * @property {Function} invoker
 * @property {string} modelId
 * @property {string} modelName
 */

const greeting = new ScenarioOutput(
  "greeting",
  "Welcome to the Amazon Bedrock Runtime client demo!",
  { header: true },
);

const selectModel = new ScenarioInput("model", "First, select a model:", {
  type: "select",
  choices: Object.values(FoundationModels).map((model) => ({
    name: model.modelName,
    value: model,
  })),
});

const enterPrompt = new ScenarioInput("prompt", "Now, enter your prompt:", {
  type: "input",
});

const printDetails = new ScenarioOutput(
  "print details",
  /**
   * @param {{ model: ModelConfig, prompt: string }} c
   */
  (c) => console.log(`Invoking ${c.model.modelName} with '${c.prompt}'...`),
);

const invokeModel = new ScenarioAction(
  "invoke model",
  /**
   * @param {{ model: ModelConfig, prompt: string, response: string }} c
   */
  async (c) => {
    const modelModule = await c.model.module();
    const invoker = c.model.invoker(modelModule);
    c.response = await invoker(c.prompt, c.model.modelId);
  },
);

const printResponse = new ScenarioOutput(
  "print response",
  /**
   * @param {{ response: string }} c
   */
  (c) => c.response,
);

const scenario = new Scenario("Amazon Bedrock Runtime Demo", [
  greeting,
  selectModel,
  enterPrompt,
  printDetails,
  invokeModel,
  printResponse,
]);

if (process.argv[1] === fileURLToPath(import.meta.url)) {
  scenario.run();
}
```
+ 如需 API 詳細資訊，請參閱《*適用於 JavaScript 的 AWS SDK API 參考*》中的下列主題。
  + [InvokeModel](https://docs.aws.amazon.com/AWSJavaScriptSDK/v3/latest/client/bedrock-runtime/command/InvokeModelCommand)
  + [InvokeModelWithResponseStream](https://docs.aws.amazon.com/AWSJavaScriptSDK/v3/latest/client/bedrock-runtime/command/InvokeModelWithResponseStreamCommand)

------
#### [ PHP ]

**適用於 PHP 的 SDK**  
 GitHub 上提供更多範例。尋找完整範例，並了解如何在 [AWS 程式碼範例儲存庫](https://github.com/awsdocs/aws-doc-sdk-examples/tree/main/php/example_code/bedrock-runtime/#code-examples)中設定和執行。
在 Amazon Bedrock 上調用多個 LLM。  

```
namespace BedrockRuntime;

class GettingStartedWithBedrockRuntime
{
    protected BedrockRuntimeService $bedrockRuntimeService;
    public function runExample()
    {
        echo "\n";
        echo "---------------------------------------------------------------------\n";
        echo "Welcome to the Amazon Bedrock Runtime getting started demo using PHP!\n";
        echo "---------------------------------------------------------------------\n";
        $bedrockRuntimeService = new BedrockRuntimeService();
        $prompt = 'In one paragraph, who are you?';
        echo "\nPrompt: " . $prompt;
        echo "\n\nAnthropic Claude:\n";
        echo $bedrockRuntimeService->invokeClaude($prompt);
        echo "\n---------------------------------------------------------------------\n";
        $image_prompt = 'stylized picture of a cute old steampunk robot';
        echo "\nImage prompt: " . $image_prompt;
        echo "\n\nStability.ai Stable Diffusion XL:\n";
        $diffusionSeed = rand(0, 4294967295);
        $style_preset = 'photographic';
        $base64 = $bedrockRuntimeService->invokeStableDiffusion($image_prompt, $diffusionSeed, $style_preset);
        $image_path = $this->saveImage($base64, 'stability.stable-diffusion-xl');
        echo "The generated image has been saved to $image_path";
        echo "\n\nAmazon Titan Image Generation:\n";
        $titanSeed = rand(0, 2147483647);
        $base64 = $bedrockRuntimeService->invokeTitanImage($image_prompt, $titanSeed);
        $image_path = $this->saveImage($base64, 'amazon.titan-image-generator-v2');
        echo "The generated image has been saved to $image_path";
    }

    private function saveImage($base64_image_data, $model_id): string
    {
        $output_dir = "output";
        if (!file_exists($output_dir)) {
            mkdir($output_dir);
        }

        $i = 1;
        while (file_exists("$output_dir/$model_id" . '_' . "$i.png")) {
            $i++;
        }

        $image_data = base64_decode($base64_image_data);
        $file_path = "$output_dir/$model_id" . '_' . "$i.png";
        $file = fopen($file_path, 'wb');
        fwrite($file, $image_data);
        fclose($file);
        return $file_path;
    }
}
```
+ 如需 API 詳細資訊，請參閱《*適用於 PHP 的 AWS SDK API 參考*》中的下列主題。
  + [InvokeModel](https://docs.aws.amazon.com/goto/SdkForPHPV3/bedrock-runtime-2023-09-30/InvokeModel)
  + [InvokeModelWithResponseStream](https://docs.aws.amazon.com/goto/SdkForPHPV3/bedrock-runtime-2023-09-30/InvokeModelWithResponseStream)

------

如需 AWS SDK 開發人員指南和程式碼範例的完整清單，請參閱 [搭配 AWS SDK 使用 Amazon Bedrock](sdk-general-information-section.md)。此主題也包含有關入門的資訊和舊版 SDK 的詳細資訊。

# 使用 Amazon Bedrock 和 Step Functions，建置和協調生成式 AI 應用程式
<a name="bedrock-runtime_example_cross_ServerlessPromptChaining_section"></a>

下列程式碼範例示範如何使用 Amazon Bedrock 和 Step Functions，建置和協調生成式 AI 應用程式。

------
#### [ Python ]

**適用於 Python 的 SDK (Boto3)**  
 Amazon Bedrock Serverless 提示鏈接案例展示，[AWS Step Functions](https://docs.aws.amazon.com/step-functions/latest/dg/welcome.html)、[Amazon Bedrock](https://docs.aws.amazon.com/bedrock/latest/userguide/what-is-bedrock.html) 和 [https://docs.aws.amazon.com/bedrock/latest/userguide/agents.html](https://docs.aws.amazon.com/bedrock/latest/userguide/agents.html) 如何用於建置和協調複雜、無伺服器和可高度擴展的生成式 AI 應用程式。該案例包含下列工作範例：  
+  針對文獻部落格撰寫指定小說的分析。此範例描述簡易、循序的提示鏈。
+  產生有關指定主題的簡短故事。此範例描述 AI 如何反覆處理先前產生的項目清單。
+  建立前往指定目的地的週末假期行程。此範例描述如何平行處理多個不同的提示。
+  向擔任電影製片的人類使用者推銷電影創意。此範例描述如何使用不同的推論參數平行處理相同的提示、如何回溯到鏈接的上一個步驟，以及如何將人工輸入包含在工作流程中。
+  根據使用者手上的配料來規劃用餐。此範例描述提示鏈如何整合兩個不同的 AI 對話，其中兩個 AI 角色互相爭論以改善最終結果。
+  尋找並總結目前最熱門的 GitHub 儲存庫。此範例說明鏈接多個與外部 API 互動的 AI 代理程式。
 如需完整的原始碼，以及有關如何設定及執行的指示，請參閱 [GitHub](https://github.com/aws-samples/amazon-bedrock-serverless-prompt-chaining) 上的完整專案。  

**此範例中使用的服務**
+ Amazon Bedrock
+ Amazon Bedrock 執行時期
+ Amazon Bedrock 代理程式
+ Amazon Bedrock 代理程式執行時期
+ 步驟函數

------

如需 AWS SDK 開發人員指南和程式碼範例的完整清單，請參閱 [搭配 AWS SDK 使用 Amazon Bedrock](sdk-general-information-section.md)。此主題也包含有關入門的資訊和舊版 SDK 的詳細資訊。

# 工具使用範例描述如何將 Amazon Bedrock 上的 AI 模型連接至自訂工具或 API
<a name="bedrock-runtime_example_bedrock-runtime_Scenario_ToolUse_section"></a>

下列程式碼範例示範如何在應用程式、生成式 AI 模型和連線工具或 API 之間建立典型的互動，以媒介 AI 與外部世界之間的互動。其使用將外部天氣 API 連接至 AI 模型的範例，以根據使用者輸入提供即時天氣資訊。

------
#### [ .NET ]

**適用於 .NET 的 SDK**  
 GitHub 上提供更多範例。尋找完整範例，並了解如何在 [AWS 程式碼範例儲存庫](https://github.com/awsdocs/aws-doc-sdk-examples/tree/main/dotnetv3/Bedrock-runtime/Scenarios/ConverseToolScenario#code-examples)中設定和執行。
案例流程的主要執行。此案例會協調使用者、Amazon Bedrock Converse API 和天氣工具之間的對話。  

```
using Amazon;
using Amazon.BedrockRuntime;
using Amazon.BedrockRuntime.Model;
using Amazon.Runtime.Documents;
using Microsoft.Extensions.DependencyInjection;
using Microsoft.Extensions.DependencyInjection.Extensions;
using Microsoft.Extensions.Hosting;
using Microsoft.Extensions.Http;
using Microsoft.Extensions.Logging;
using Microsoft.Extensions.Logging.Console;

namespace ConverseToolScenario;

public static class ConverseToolScenario
{
    /*
    Before running this .NET code example, set up your development environment, including your credentials.

    This demo illustrates a tool use scenario using Amazon Bedrock's Converse API and a weather tool.
    The script interacts with a foundation model on Amazon Bedrock to provide weather information based on user
    input. It uses the Open-Meteo API (https://open-meteo.com) to retrieve current weather data for a given location.
   */

    public static BedrockActionsWrapper _bedrockActionsWrapper = null!;
    public static WeatherTool _weatherTool = null!;
    public static bool _interactive = true;

    // Change this string to use a different model with Converse API.
    private static string model_id = "amazon.nova-lite-v1:0";

    private static string system_prompt = @"
        You are a weather assistant that provides current weather data for user-specified locations using only
        the Weather_Tool, which expects latitude and longitude. Infer the coordinates from the location yourself.
        If the user specifies a state, country, or region, infer the locations of cities within that state.
        If the user provides coordinates, infer the approximate location and refer to it in your response.
        To use the tool, you strictly apply the provided tool specification.

        - Explain your step-by-step process, and give brief updates before each step.
        - Only use the Weather_Tool for data. Never guess or make up information. 
        - Repeat the tool use for subsequent requests if necessary.
        - If the tool errors, apologize, explain weather is unavailable, and suggest other options.
        - Report temperatures in °C (°F) and wind in km/h (mph). Keep weather reports concise. Sparingly use
          emojis where appropriate.
        - Only respond to weather queries. Remind off-topic users of your purpose. 
        - Never claim to search online, access external data, or use tools besides Weather_Tool.
        - Complete the entire process until you have all required data before sending the complete response.
    "
    ;

    private static string default_prompt = "What is the weather like in Seattle?";

    // The maximum number of recursive calls allowed in the tool use function.
    // This helps prevent infinite loops and potential performance issues.
    private static int max_recursions = 5;

    public static async Task Main(string[] args)
    {
        // Set up dependency injection for the Amazon service.
        using var host = Host.CreateDefaultBuilder(args)
            .ConfigureLogging(logging =>
                logging.AddFilter("System", LogLevel.Error)
                    .AddFilter<ConsoleLoggerProvider>("Microsoft", LogLevel.Trace))
            .ConfigureServices((_, services) =>
                services.AddHttpClient()
                    .AddSingleton<IAmazonBedrockRuntime>(_ => new AmazonBedrockRuntimeClient(RegionEndpoint.USEast1)) // Specify a region that has access to the chosen model.
                    .AddTransient<BedrockActionsWrapper>()
                    .AddTransient<WeatherTool>()
                    .RemoveAll<IHttpMessageHandlerBuilderFilter>()
            )
            .Build();

        ServicesSetup(host);

        try
        {
            await RunConversationAsync();

        }
        catch (Exception ex)
        {
            Console.WriteLine(new string('-', 80));
            Console.WriteLine($"There was a problem running the scenario: {ex.Message}");
            Console.WriteLine(new string('-', 80));
        }
        finally
        {
            Console.WriteLine(
                "Amazon Bedrock Converse API with Tool Use Feature Scenario is complete.");
            Console.WriteLine(new string('-', 80));
        }
    }

    /// <summary>
    /// Populate the services for use within the console application.
    /// </summary>
    /// <param name="host">The services host.</param>
    private static void ServicesSetup(IHost host)
    {
        _bedrockActionsWrapper = host.Services.GetRequiredService<BedrockActionsWrapper>();
        _weatherTool = host.Services.GetRequiredService<WeatherTool>();
    }

    /// <summary>
    /// Starts the conversation with the user and handles the interaction with Bedrock.
    /// </summary>
    /// <returns>The conversation array.</returns>
    public static async Task<List<Message>> RunConversationAsync()
    {
        // Print the greeting and a short user guide
        PrintHeader();

        // Start with an empty conversation
        var conversation = new List<Message>();

        // Get the first user input
        var userInput = await GetUserInputAsync();

        while (userInput != null)
        {
            // Create a new message with the user input and append it to the conversation
            var message = new Message { Role = ConversationRole.User, Content = new List<ContentBlock> { new ContentBlock { Text = userInput } } };
            conversation.Add(message);

            // Send the conversation to Amazon Bedrock
            var bedrockResponse = await SendConversationToBedrock(conversation);

            // Recursively handle the model's response until the model has returned its final response or the recursion counter has reached 0
            await ProcessModelResponseAsync(bedrockResponse, conversation, max_recursions);

            // Repeat the loop until the user decides to exit the application
            userInput = await GetUserInputAsync();
        }

        PrintFooter();
        return conversation;
    }

    /// <summary>
    /// Sends the conversation, the system prompt, and the tool spec to Amazon Bedrock, and returns the response.
    /// </summary>
    /// <param name="conversation">The conversation history including the next message to send.</param>
    /// <returns>The response from Amazon Bedrock.</returns>
    private static async Task<ConverseResponse> SendConversationToBedrock(List<Message> conversation)
    {
        Console.WriteLine("\tCalling Bedrock...");

        // Send the conversation, system prompt, and tool configuration, and return the response
        return await _bedrockActionsWrapper.SendConverseRequestAsync(model_id, system_prompt, conversation, _weatherTool.GetToolSpec());
    }

    /// <summary>
    /// Processes the response received via Amazon Bedrock and performs the necessary actions based on the stop reason.
    /// </summary>
    /// <param name="modelResponse">The model's response returned via Amazon Bedrock.</param>
    /// <param name="conversation">The conversation history.</param>
    /// <param name="maxRecursion">The maximum number of recursive calls allowed.</param>
    private static async Task ProcessModelResponseAsync(ConverseResponse modelResponse, List<Message> conversation, int maxRecursion)
    {
        if (maxRecursion <= 0)
        {
            // Stop the process, the number of recursive calls could indicate an infinite loop
            Console.WriteLine("\tWarning: Maximum number of recursions reached. Please try again.");
        }

        // Append the model's response to the ongoing conversation
        conversation.Add(modelResponse.Output.Message);

        if (modelResponse.StopReason == "tool_use")
        {
            // If the stop reason is "tool_use", forward everything to the tool use handler
            await HandleToolUseAsync(modelResponse.Output, conversation, maxRecursion - 1);
        }

        if (modelResponse.StopReason == "end_turn")
        {
            // If the stop reason is "end_turn", print the model's response text, and finish the process
            PrintModelResponse(modelResponse.Output.Message.Content[0].Text);
            if (!_interactive)
            {
                default_prompt = "x";
            }
        }
    }

    /// <summary>
    /// Handles the tool use case by invoking the specified tool and sending the tool's response back to Bedrock.
    /// The tool response is appended to the conversation, and the conversation is sent back to Amazon Bedrock for further processing.
    /// </summary>
    /// <param name="modelResponse">The model's response containing the tool use request.</param>
    /// <param name="conversation">The conversation history.</param>
    /// <param name="maxRecursion">The maximum number of recursive calls allowed.</param>
    public static async Task HandleToolUseAsync(ConverseOutput modelResponse, List<Message> conversation, int maxRecursion)
    {
        // Initialize an empty list of tool results
        var toolResults = new List<ContentBlock>();

        // The model's response can consist of multiple content blocks
        foreach (var contentBlock in modelResponse.Message.Content)
        {
            if (!String.IsNullOrEmpty(contentBlock.Text))
            {
                // If the content block contains text, print it to the console
                PrintModelResponse(contentBlock.Text);
            }

            if (contentBlock.ToolUse != null)
            {
                // If the content block is a tool use request, forward it to the tool
                var toolResponse = await InvokeTool(contentBlock.ToolUse);

                // Add the tool use ID and the tool's response to the list of results
                toolResults.Add(new ContentBlock
                {
                    ToolResult = new ToolResultBlock()
                    {
                        ToolUseId = toolResponse.ToolUseId,
                        Content = new List<ToolResultContentBlock>()
                            { new ToolResultContentBlock { Json = toolResponse.Content } }
                    }
                });
            }
        }

        // Embed the tool results in a new user message
        var message = new Message() { Role = ConversationRole.User, Content = toolResults };

        // Append the new message to the ongoing conversation
        conversation.Add(message);

        // Send the conversation to Amazon Bedrock
        var response = await SendConversationToBedrock(conversation);

        // Recursively handle the model's response until the model has returned its final response or the recursion counter has reached 0
        await ProcessModelResponseAsync(response, conversation, maxRecursion);
    }

    /// <summary>
    /// Invokes the specified tool with the given payload and returns the tool's response.
    /// If the requested tool does not exist, an error message is returned.
    /// </summary>
    /// <param name="payload">The payload containing the tool name and input data.</param>
    /// <returns>The tool's response or an error message.</returns>
    public static async Task<ToolResponse> InvokeTool(ToolUseBlock payload)
    {
        var toolName = payload.Name;

        if (toolName == "Weather_Tool")
        {
            var inputData = payload.Input.AsDictionary();
            PrintToolUse(toolName, inputData);

            // Invoke the weather tool with the input data provided
            var weatherResponse = await _weatherTool.FetchWeatherDataAsync(inputData["latitude"].ToString(), inputData["longitude"].ToString());
            return new ToolResponse { ToolUseId = payload.ToolUseId, Content = weatherResponse };
        }
        else
        {
            var errorMessage = $"\tThe requested tool with name '{toolName}' does not exist.";
            return new ToolResponse { ToolUseId = payload.ToolUseId, Content = new { error = true, message = errorMessage } };
        }
    }


    /// <summary>
    /// Prompts the user for input and returns the user's response.
    /// Returns null if the user enters 'x' to exit.
    /// </summary>
    /// <param name="prompt">The prompt to display to the user.</param>
    /// <returns>The user's input or null if the user chooses to exit.</returns>
    private static async Task<string?> GetUserInputAsync(string prompt = "\tYour weather info request:")
    {
        var userInput = default_prompt;
        if (_interactive)
        {
            Console.WriteLine(new string('*', 80));
            Console.WriteLine($"{prompt} (x to exit): \n\t");
            userInput = Console.ReadLine();
        }

        if (string.IsNullOrWhiteSpace(userInput))
        {
            prompt = "\tPlease enter your weather info request, e.g. the name of a city";
            return await GetUserInputAsync(prompt);
        }

        if (userInput.ToLowerInvariant() == "x")
        {
            return null;
        }

        return userInput;
    }

    /// <summary>
    /// Logs the welcome message and usage guide for the tool use demo.
    /// </summary>
    public static void PrintHeader()
    {
        Console.WriteLine(@"
        =================================================
        Welcome to the Amazon Bedrock Tool Use demo!
        =================================================

        This assistant provides current weather information for user-specified locations.
        You can ask for weather details by providing the location name or coordinates. Weather information
        will be provided using a custom Tool and open-meteo API.

        Example queries:
        - What's the weather like in New York?
        - Current weather for latitude 40.70, longitude -74.01
        - Is it warmer in Rome or Barcelona today?

        To exit the program, simply type 'x' and press Enter.

        P.S.: You're not limited to single locations, or even to using English!
        Have fun and experiment with the app!
        ");
    }

    /// <summary>
    /// Logs the footer information for the tool use demo.
    /// </summary>
    public static void PrintFooter()
    {
        Console.WriteLine(@"
        =================================================
        Thank you for checking out the Amazon Bedrock Tool Use demo. We hope you
        learned something new, or got some inspiration for your own apps today!

        For more Bedrock examples in different programming languages, have a look at:
        https://docs.aws.amazon.com/bedrock/latest/userguide/service_code_examples.html
        =================================================
        ");
    }

    /// <summary>
    /// Logs information about the tool use.
    /// </summary>
    /// <param name="toolName">The name of the tool being used.</param>
    /// <param name="inputData">The input data for the tool.</param>
    public static void PrintToolUse(string toolName, Dictionary<string, Document> inputData)
    {
        Console.WriteLine($"\n\tInvoking tool: {toolName} with input: {inputData["latitude"].ToString()}, {inputData["longitude"].ToString()}...\n");
    }

    /// <summary>
    /// Logs the model's response.
    /// </summary>
    /// <param name="message">The model's response message.</param>
    public static void PrintModelResponse(string message)
    {
        Console.WriteLine("\tThe model's response:\n");
        Console.WriteLine(message);
        Console.WriteLine();
    }
}
```
示範時使用的天氣工具。此檔案定義工具規格，並實作邏輯，以從 Open-Meteo API 用來擷取天氣資料。  

```
using Amazon.BedrockRuntime.Model;
using Amazon.Runtime.Documents;
using Microsoft.Extensions.Logging;

namespace ConverseToolScenario;

/// <summary>
/// Weather tool that will be invoked when requested by the Bedrock response.
/// </summary>
public class WeatherTool
{
    private readonly ILogger<WeatherTool> _logger;
    private readonly IHttpClientFactory _httpClientFactory;

    public WeatherTool(ILogger<WeatherTool> logger, IHttpClientFactory httpClientFactory)
    {
        _logger = logger;
        _httpClientFactory = httpClientFactory;
    }

    /// <summary>
    /// Returns the JSON Schema specification for the Weather tool. The tool specification
    /// defines the input schema and describes the tool's functionality.
    /// For more information, see https://json-schema.org/understanding-json-schema/reference.
    /// </summary>
    /// <returns>The tool specification for the Weather tool.</returns>
    public ToolSpecification GetToolSpec()
    {
        ToolSpecification toolSpecification = new ToolSpecification();

        toolSpecification.Name = "Weather_Tool";
        toolSpecification.Description = "Get the current weather for a given location, based on its WGS84 coordinates.";

        Document toolSpecDocument = Document.FromObject(
            new
            {
                type = "object",
                properties = new
                {
                    latitude = new
                    {
                        type = "string",
                        description = "Geographical WGS84 latitude of the location."
                    },
                    longitude = new
                    {
                        type = "string",
                        description = "Geographical WGS84 longitude of the location."
                    }
                },
                required = new[] { "latitude", "longitude" }
            });

        toolSpecification.InputSchema = new ToolInputSchema() { Json = toolSpecDocument };
        return toolSpecification;
    }

    /// <summary>
    /// Fetches weather data for the given latitude and longitude using the Open-Meteo API.
    /// Returns the weather data or an error message if the request fails.
    /// </summary>
    /// <param name="latitude">The latitude of the location.</param>
    /// <param name="longitude">The longitude of the location.</param>
    /// <returns>The weather data or an error message.</returns>
    public async Task<Document> FetchWeatherDataAsync(string latitude, string longitude)
    {
        string endpoint = "https://api.open-meteo.com/v1/forecast";

        try
        {
            var httpClient = _httpClientFactory.CreateClient();
            var response = await httpClient.GetAsync($"{endpoint}?latitude={latitude}&longitude={longitude}&current_weather=True");
            response.EnsureSuccessStatusCode();
            var weatherData = await response.Content.ReadAsStringAsync();

            Document weatherDocument = Document.FromObject(
                new { weather_data = weatherData });

            return weatherDocument;
        }
        catch (HttpRequestException e)
        {
            _logger.LogError(e, "Error fetching weather data: {Message}", e.Message);
            throw;
        }
        catch (Exception e)
        {
            _logger.LogError(e, "Unexpected error fetching weather data: {Message}", e.Message);
            throw;
        }
    }
}
```
具有工具組態的 Converse API 動作。  

```
/// <summary>
/// Wrapper class for interacting with the Amazon Bedrock Converse API.
/// </summary>
public class BedrockActionsWrapper
{
    private readonly IAmazonBedrockRuntime _bedrockClient;
    private readonly ILogger<BedrockActionsWrapper> _logger;

    /// <summary>
    /// Initializes a new instance of the <see cref="BedrockActionsWrapper"/> class.
    /// </summary>
    /// <param name="bedrockClient">The Bedrock Converse API client.</param>
    /// <param name="logger">The logger instance.</param>
    public BedrockActionsWrapper(IAmazonBedrockRuntime bedrockClient, ILogger<BedrockActionsWrapper> logger)
    {
        _bedrockClient = bedrockClient;
        _logger = logger;
    }

    /// <summary>
    /// Sends a Converse request to the Amazon Bedrock Converse API.
    /// </summary>
    /// <param name="modelId">The Bedrock Model Id.</param>
    /// <param name="systemPrompt">A system prompt instruction.</param>
    /// <param name="conversation">The array of messages in the conversation.</param>
    /// <param name="toolSpec">The specification for a tool.</param>
    /// <returns>The response of the model.</returns>
    public async Task<ConverseResponse> SendConverseRequestAsync(string modelId, string systemPrompt, List<Message> conversation, ToolSpecification toolSpec)
    {
        try
        {
            var request = new ConverseRequest()
            {
                ModelId = modelId,
                System = new List<SystemContentBlock>()
                {
                    new SystemContentBlock()
                    {
                        Text = systemPrompt
                    }
                },
                Messages = conversation,
                ToolConfig = new ToolConfiguration()
                {
                    Tools = new List<Tool>()
                    {
                        new Tool()
                        {
                            ToolSpec = toolSpec
                        }
                    }
                }
            };

            var response = await _bedrockClient.ConverseAsync(request);

            return response;
        }
        catch (ModelNotReadyException ex)
        {
            _logger.LogError(ex, "Model not ready, please wait and try again.");
            throw;
        }
        catch (AmazonBedrockRuntimeException ex)
        {
            _logger.LogError(ex, "Error occurred while sending Converse request.");
            throw;
        }
    }
}
```
+  如需 API 詳細資訊，請參閱《適用於 .NET 的 AWS SDK API 參考》**中的 [Converse](https://docs.aws.amazon.com/goto/DotNetSDKV3/bedrock-runtime-2023-09-30/Converse)。

------
#### [ Java ]

**SDK for Java 2.x**  
 GitHub 上提供更多範例。尋找完整範例，並了解如何在 [AWS 程式碼範例儲存庫](https://github.com/awsdocs/aws-doc-sdk-examples/tree/main/javav2/example_code/bedrock-runtime#code-examples)中設定和執行。
案例流程的主要執行。此案例會協調使用者、Amazon Bedrock Converse API 和天氣工具之間的對話。  

```
/*
 This demo illustrates a tool use scenario using Amazon Bedrock's Converse API and a weather tool.
 The program interacts with a foundation model on Amazon Bedrock to provide weather information based on user
 input. It uses the Open-Meteo API (https://open-meteo.com) to retrieve current weather data for a given location.
 */
public class BedrockScenario {
    public static final String DASHES = new String(new char[80]).replace("\0", "-");
    private static String modelId = "amazon.nova-lite-v1:0";
    private static String defaultPrompt = "What is the weather like in Seattle?";
    private static WeatherTool weatherTool = new WeatherTool();

    // The maximum number of recursive calls allowed in the tool use function.
    // This helps prevent infinite loops and potential performance issues.
    private static int maxRecursions = 5;
    static BedrockActions bedrockActions = new BedrockActions();
    public static boolean interactive = true;

    private static final String systemPrompt = """
            You are a weather assistant that provides current weather data for user-specified locations using only
            the Weather_Tool, which expects latitude and longitude. Infer the coordinates from the location yourself.
            If the user provides coordinates, infer the approximate location and refer to it in your response.
            To use the tool, you strictly apply the provided tool specification.
            
            - Explain your step-by-step process, and give brief updates before each step.
            - Only use the Weather_Tool for data. Never guess or make up information. 
            - Repeat the tool use for subsequent requests if necessary.
            - If the tool errors, apologize, explain weather is unavailable, and suggest other options.
            - Report temperatures in °C (°F) and wind in km/h (mph). Keep weather reports concise. Sparingly use
              emojis where appropriate.
            - Only respond to weather queries. Remind off-topic users of your purpose. 
            - Never claim to search online, access external data, or use tools besides Weather_Tool.
            - Complete the entire process until you have all required data before sending the complete response.
            """;

    public static void main(String[] args) {
        Scanner scanner = new Scanner(System.in);
        System.out.println("""
                =================================================
                Welcome to the Amazon Bedrock Tool Use demo!
                =================================================
                
                This assistant provides current weather information for user-specified locations.
                You can ask for weather details by providing the location name or coordinates.
                
                Example queries:
                - What's the weather like in New York?
                - Current weather for latitude 40.70, longitude -74.01
                - Is it warmer in Rome or Barcelona today?
                
                To exit the program, simply type 'x' and press Enter.
                
                P.S.: You're not limited to single locations, or even to using English!
                Have fun and experiment with the app!
                """);
        System.out.println(DASHES);

        try {
            runConversation(scanner);

        } catch (Exception ex) {
            System.out.println("There was a problem running the scenario: " + ex.getMessage());
        }

        waitForInputToContinue(scanner);

        System.out.println(DASHES);
        System.out.println("Amazon Bedrock Converse API with Tool Use Feature Scenario is complete.");
        System.out.println(DASHES);
    }

    /**
     * Starts the conversation with the user and handles the interaction with Bedrock.
     */
    private static List<Message> runConversation(Scanner scanner) {
        List<Message> conversation = new ArrayList<>();

        // Get the first user input
        String userInput = getUserInput("Your weather info request:", scanner);
        System.out.println(userInput);

        while (userInput != null) {
            ContentBlock block = ContentBlock.builder()
                    .text(userInput)
                    .build();

            List<ContentBlock> blockList = new ArrayList<>();
            blockList.add(block);

            Message message = Message.builder()
                    .role(ConversationRole.USER)
                    .content(blockList)
                    .build();

            conversation.add(message);

            // Send the conversation to Amazon Bedrock.
            ConverseResponse bedrockResponse = sendConversationToBedrock(conversation);

            // Recursively handle the model's response until the model has returned its final response or the recursion counter has reached 0.
            processModelResponse(bedrockResponse, conversation, maxRecursions);

            // Repeat the loop until the user decides to exit the application.
            userInput = getUserInput("Your weather info request:", scanner);
        }
        printFooter();
        return conversation;
    }

    /**
     * Processes the response from the model and updates the conversation accordingly.
     *
     * @param modelResponse the response from the model
     * @param conversation  the ongoing conversation
     * @param maxRecursion  the maximum number of recursions allowed
     */
    private static void processModelResponse(ConverseResponse modelResponse, List<Message> conversation, int maxRecursion) {
        if (maxRecursion <= 0) {
            // Stop the process, the number of recursive calls could indicate an infinite loop
            System.out.println("\tWarning: Maximum number of recursions reached. Please try again.");
        }

        // Append the model's response to the ongoing conversation
        conversation.add(modelResponse.output().message());

        String modelResponseVal = modelResponse.stopReasonAsString();
        if (modelResponseVal.compareTo("tool_use") == 0) {
            // If the stop reason is "tool_use", forward everything to the tool use handler
            handleToolUse(modelResponse.output(), conversation, maxRecursion - 1);
        }

        if (modelResponseVal.compareTo("end_turn") == 0) {
            // If the stop reason is "end_turn", print the model's response text, and finish the process
            PrintModelResponse(modelResponse.output().message().content().get(0).text());
            if (!interactive) {
                defaultPrompt = "x";
            }
        }
    }

    /**
     * Handles the use of a tool by the model in a conversation.
     *
     * @param modelResponse the response from the model, which may include a tool use request
     * @param conversation  the current conversation, which will be updated with the tool use results
     * @param maxRecursion  the maximum number of recursive calls allowed to handle the model's response
     */
    private static void handleToolUse(ConverseOutput modelResponse, List<Message> conversation, int maxRecursion) {
        List<ContentBlock> toolResults = new ArrayList<>();

        // The model's response can consist of multiple content blocks
        for (ContentBlock contentBlock : modelResponse.message().content()) {
            if (contentBlock.text() != null && !contentBlock.text().isEmpty()) {
                // If the content block contains text, print it to the console
                PrintModelResponse(contentBlock.text());
            }

            if (contentBlock.toolUse() != null) {
                ToolResponse toolResponse = invokeTool(contentBlock.toolUse());

                // Add the tool use ID and the tool's response to the list of results
                List<ToolResultContentBlock> contentBlockList = new ArrayList<>();
                ToolResultContentBlock block = ToolResultContentBlock.builder()
                        .json(toolResponse.getContent())
                        .build();
                contentBlockList.add(block);

                ToolResultBlock toolResultBlock = ToolResultBlock.builder()
                        .toolUseId(toolResponse.getToolUseId())
                        .content(contentBlockList)
                        .build();

                ContentBlock contentBlock1 = ContentBlock.builder()
                        .toolResult(toolResultBlock)
                        .build();

                toolResults.add(contentBlock1);
            }
        }

        // Embed the tool results in a new user message
        Message message = Message.builder()
                .role(ConversationRole.USER)
                .content(toolResults)
                .build();

        // Append the new message to the ongoing conversation
        //conversation.add(message);
        conversation.add(message);

        // Send the conversation to Amazon Bedrock
        var response = sendConversationToBedrock(conversation);

        // Recursively handle the model's response until the model has returned its final response or the recursion counter has reached 0
        processModelResponse(response, conversation, maxRecursion);
    }

    // Invokes the specified tool with the given payload and returns the tool's response.
    // If the requested tool does not exist, an error message is returned.
    private static ToolResponse invokeTool(ToolUseBlock payload) {
        String toolName = payload.name();

        if (Objects.equals(toolName, "Weather_Tool")) {
            Map<String, Document> inputData = payload.input().asMap();
            printToolUse(toolName, inputData);

            // Invoke the weather tool with the input data provided
            Document weatherResponse = weatherTool.fetchWeatherData(inputData.get("latitude").toString(), inputData.get("longitude").toString());

            ToolResponse toolResponse = new ToolResponse();
            toolResponse.setContent(weatherResponse);
            toolResponse.setToolUseId(payload.toolUseId());
            return toolResponse;
        } else {
            String errorMessage = "The requested tool with name " + toolName + " does not exist.";
            System.out.println(errorMessage);
            return null;
        }
    }

    public static void printToolUse(String toolName, Map<String, Document> inputData) {
        System.out.println("Invoking tool: " + toolName + " with input: " + inputData.get("latitude").toString() + ", " + inputData.get("longitude").toString() + "...");
    }

    private static void PrintModelResponse(String message) {
        System.out.println("\tThe model's response:\n");
        System.out.println(message);
        System.out.println("");
    }

    private static ConverseResponse sendConversationToBedrock(List<Message> conversation) {
        System.out.println("Calling Bedrock...");

        try {
            return bedrockActions.sendConverseRequestAsync(modelId, systemPrompt, conversation, weatherTool.getToolSpec());
        } catch (ModelNotReadyException ex) {
             System.err.println("Model is not ready. Please try again later: " + ex.getMessage());
            throw ex;
        } catch (BedrockRuntimeException ex) {
            System.err.println("Bedrock service error: " + ex.getMessage());
            throw ex;
        } catch (RuntimeException ex) {
            System.err.println("Unexpected error occurred: " + ex.getMessage());
            throw ex;
        }
    }

    private static ConverseResponse sendConversationToBedrockwithSpec(List<Message> conversation, ToolSpecification toolSpec) {
        System.out.println("Calling Bedrock...");

        // Send the conversation, system prompt, and tool configuration, and return the response
        return bedrockActions.sendConverseRequestAsync(modelId, systemPrompt, conversation, toolSpec);
    }

    public static String getUserInput(String prompt, Scanner scanner) {
        String userInput = defaultPrompt;
        if (interactive) {
            System.out.println("*".repeat(80));
            System.out.println(prompt + " (x to exit): \n\t");
            userInput = scanner.nextLine();
        }

        if (userInput == null || userInput.trim().isEmpty()) {
            return getUserInput("\tPlease enter your weather info request, e.g., the name of a city", scanner);
        }

        if (userInput.equalsIgnoreCase("x")) {
            return null;
        }

        return userInput;
    }

    private static void waitForInputToContinue(Scanner scanner) {
        while (true) {
            System.out.println("");
            System.out.println("Enter 'c' followed by <ENTER> to continue:");
            String input = scanner.nextLine();

            if (input.trim().equalsIgnoreCase("c")) {
                System.out.println("Continuing with the program...");
                System.out.println("");
                break;
            } else {
                // Handle invalid input.
                System.out.println("Invalid input. Please try again.");
            }
        }
    }

    public static void printFooter() {
        System.out.println("""
                =================================================
                Thank you for checking out the Amazon Bedrock Tool Use demo. We hope you
                learned something new, or got some inspiration for your own apps today!
                
                For more Bedrock examples in different programming languages, have a look at:
                https://docs.aws.amazon.com/bedrock/latest/userguide/service_code_examples.html
                =================================================
                """);
    }
}
```
示範時使用的天氣工具。此檔案定義工具規格，並實作邏輯，以從 Open-Meteo API 用來擷取天氣資料。  

```
public class WeatherTool {

    private static final Logger logger = LoggerFactory.getLogger(WeatherTool.class);
    private static java.net.http.HttpClient httpClient = null;

    /**
     * Returns the JSON Schema specification for the Weather tool. The tool specification
     * defines the input schema and describes the tool's functionality.
     * For more information, see https://json-schema.org/understanding-json-schema/reference.
     *
     * @return The tool specification for the Weather tool.
     */
    public ToolSpecification getToolSpec() {
        Map<String, Document> latitudeMap = new HashMap<>();
        latitudeMap.put("type", Document.fromString("string"));
        latitudeMap.put("description", Document.fromString("Geographical WGS84 latitude of the location."));

        // Create the nested "longitude" object
        Map<String, Document> longitudeMap = new HashMap<>();
        longitudeMap.put("type", Document.fromString("string"));
        longitudeMap.put("description", Document.fromString("Geographical WGS84 longitude of the location."));

        // Create the "properties" object
        Map<String, Document> propertiesMap = new HashMap<>();
        propertiesMap.put("latitude", Document.fromMap(latitudeMap));
        propertiesMap.put("longitude", Document.fromMap(longitudeMap));

        // Create the "required" array
        List<Document> requiredList = new ArrayList<>();
        requiredList.add(Document.fromString("latitude"));
        requiredList.add(Document.fromString("longitude"));

        // Create the root object
        Map<String, Document> rootMap = new HashMap<>();
        rootMap.put("type", Document.fromString("object"));
        rootMap.put("properties", Document.fromMap(propertiesMap));
        rootMap.put("required", Document.fromList(requiredList));

        // Now create the Document representing the JSON schema
        Document document = Document.fromMap(rootMap);

        ToolSpecification specification = ToolSpecification.builder()
            .name("Weather_Tool")
            .description("Get the current weather for a given location, based on its WGS84 coordinates.")
            .inputSchema(ToolInputSchema.builder()
                .json(document)
                .build())
            .build();

        return specification;
    }

    /**
     * Fetches weather data for the given latitude and longitude.
     *
     * @param latitude  the latitude coordinate
     * @param longitude the longitude coordinate
     * @return a {@link CompletableFuture} containing the weather data as a JSON string
     */
    public Document fetchWeatherData(String latitude, String longitude) {
        HttpClient httpClient = HttpClient.newHttpClient();

        // Ensure no extra double quotes
        latitude = latitude.replace("\"", "");
        longitude = longitude.replace("\"", "");

        String endpoint = "https://api.open-meteo.com/v1/forecast";
        String url = String.format("%s?latitude=%s&longitude=%s&current_weather=True", endpoint, latitude, longitude);

        HttpRequest request = HttpRequest.newBuilder()
            .uri(URI.create(url))
            .build();

        try {
            HttpResponse<String> response = httpClient.send(request, HttpResponse.BodyHandlers.ofString());
            if (response.statusCode() == 200) {
                String weatherJson = response.body();
                System.out.println(weatherJson);
                ObjectMapper objectMapper = new ObjectMapper();
                Map<String, Object> rawMap = objectMapper.readValue(weatherJson, new TypeReference<Map<String, Object>>() {});
                Map<String, Document> documentMap = convertToDocumentMap(rawMap);


                Document weatherDocument = Document.fromMap(documentMap);
                System.out.println(weatherDocument);
                return weatherDocument;
            } else {
                throw new RuntimeException("Error fetching weather data: " + response.statusCode());
            }
        } catch (Exception e) {
            System.out.println("Error fetching weather data: " + e.getMessage());
            throw new RuntimeException("Error fetching weather data", e);
        }

    }

    private static Map<String, Document> convertToDocumentMap(Map<String, Object> inputMap) {
        Map<String, Document> result = new HashMap<>();
        for (Map.Entry<String, Object> entry : inputMap.entrySet()) {
            result.put(entry.getKey(), convertToDocument(entry.getValue()));
        }
        return result;
    }

    // Convert different types of Objects to Document
    private static Document convertToDocument(Object value) {
        if (value instanceof Map) {
            return Document.fromMap(convertToDocumentMap((Map<String, Object>) value));
        } else if (value instanceof Integer) {
            return Document.fromNumber(SdkNumber.fromInteger((Integer) value));
        } else if (value instanceof Double) {  //
            return Document.fromNumber(SdkNumber.fromDouble((Double) value));
        } else if (value instanceof Boolean) {
            return Document.fromBoolean((Boolean) value);
        } else if (value instanceof String) {
            return Document.fromString((String) value);
        }
        return Document.fromNull(); // Handle null values safely
    }
}
```
具有工具組態的 Converse API 動作。  

```
    /**
     * Sends an asynchronous converse request to the AI model.
     *
     * @param modelId      the unique identifier of the AI model to be used for the converse request
     * @param systemPrompt the system prompt to be included in the converse request
     * @param conversation a list of messages representing the conversation history
     * @param toolSpec     the specification of the tool to be used in the converse request
     * @return the converse response received from the AI model
     */
    public ConverseResponse sendConverseRequestAsync(String modelId, String systemPrompt, List<Message> conversation, ToolSpecification toolSpec) {
        List<Tool> toolList = new ArrayList<>();
        Tool tool = Tool.builder()
            .toolSpec(toolSpec)
            .build();

        toolList.add(tool);

        ToolConfiguration configuration = ToolConfiguration.builder()
            .tools(toolList)
            .build();

        SystemContentBlock block = SystemContentBlock.builder()
            .text(systemPrompt)
            .build();

        ConverseRequest request = ConverseRequest.builder()
            .modelId(modelId)
            .system(block)
            .messages(conversation)
            .toolConfig(configuration)
            .build();

        try {
            ConverseResponse response = getClient().converse(request).join();
            return response;

        } catch (ModelNotReadyException ex) {
            throw new RuntimeException("Model is not ready: " + ex.getMessage(), ex);
        } catch (BedrockRuntimeException ex) {
            throw new RuntimeException("Failed to converse with Bedrock model: " + ex.getMessage(), ex);
        }
    }
```
+  如需 API 詳細資訊，請參閱《AWS SDK for Java 2.x API 參考》**中的 [Converse](https://docs.aws.amazon.com/goto/SdkForJavaV2/bedrock-runtime-2023-09-30/Converse)。

------
#### [ JavaScript ]

**適用於 JavaScript (v3) 的 SDK**  
 GitHub 上提供更多範例。尋找完整範例，並了解如何在 [AWS 程式碼範例儲存庫](https://github.com/awsdocs/aws-doc-sdk-examples/tree/main/javascriptv3/example_code/bedrock-runtime/scenarios/converse_tool_scenario#code-examples)中設定和執行。
案例流程的主要執行。此案例會協調使用者、Amazon Bedrock Converse API 和天氣工具之間的對話。  

```
/* Before running this JavaScript code example, set up your development environment, including your credentials.
This demo illustrates a tool use scenario using Amazon Bedrock's Converse API and a weather tool.
The script interacts with a foundation model on Amazon Bedrock to provide weather information based on user
input. It uses the Open-Meteo API (https://open-meteo.com) to retrieve current weather data for a given location.*/

import {
  Scenario,
  ScenarioAction,
  ScenarioInput,
  ScenarioOutput,
} from "@aws-doc-sdk-examples/lib/scenario/index.js";
import {
  BedrockRuntimeClient,
  ConverseCommand,
} from "@aws-sdk/client-bedrock-runtime";

import { parseArgs } from "node:util";
import { fileURLToPath } from "node:url";
import data from "./questions.json" with { type: "json" };
import toolConfig from "./tool_config.json" with { type: "json" };

const __filename = fileURLToPath(import.meta.url);

const systemPrompt = [
  {
    text:
      "You are a weather assistant that provides current weather data for user-specified locations using only\n" +
      "the Weather_Tool, which expects latitude and longitude. Infer the coordinates from the location yourself.\n" +
      "If the user provides coordinates, infer the approximate location and refer to it in your response.\n" +
      "To use the tool, you strictly apply the provided tool specification.\n" +
      "If the user specifies a state, country, or region, infer the locations of cities within that state.\n" +
      "\n" +
      "- Explain your step-by-step process, and give brief updates before each step.\n" +
      "- Only use the Weather_Tool for data. Never guess or make up information. \n" +
      "- Repeat the tool use for subsequent requests if necessary.\n" +
      "- If the tool errors, apologize, explain weather is unavailable, and suggest other options.\n" +
      "- Report temperatures in °C (°F) and wind in km/h (mph). Keep weather reports concise. Sparingly use\n" +
      "  emojis where appropriate.\n" +
      "- Only respond to weather queries. Remind off-topic users of your purpose. \n" +
      "- Never claim to search online, access external data, or use tools besides Weather_Tool.\n" +
      "- Complete the entire process until you have all required data before sending the complete response.",
  },
];
const tools_config = toolConfig;

/// Starts the conversation with the user and handles the interaction with Bedrock.
async function askQuestion(userMessage) {
  // The maximum number of recursive calls allowed in the tool use function.
  // This helps prevent infinite loops and potential performance issues.
  const max_recursions = 5;
  const messages = [
    {
      role: "user",
      content: [{ text: userMessage }],
    },
  ];
  try {
    const response = await SendConversationtoBedrock(messages);
    await ProcessModelResponseAsync(response, messages, max_recursions);
  } catch (error) {
    console.log("error ", error);
  }
}

// Sends the conversation, the system prompt, and the tool spec to Amazon Bedrock, and returns the response.
// param "messages" - The conversation history including the next message to send.
// return - The response from Amazon Bedrock.
async function SendConversationtoBedrock(messages) {
  const bedRockRuntimeClient = new BedrockRuntimeClient({
    region: "us-east-1",
  });
  try {
    const modelId = "amazon.nova-lite-v1:0";
    const response = await bedRockRuntimeClient.send(
      new ConverseCommand({
        modelId: modelId,
        messages: messages,
        system: systemPrompt,
        toolConfig: tools_config,
      }),
    );
    return response;
  } catch (caught) {
    if (caught.name === "ModelNotReady") {
      console.log(
        "`${caught.name}` - Model not ready, please wait and try again.",
      );
      throw caught;
    }
    if (caught.name === "BedrockRuntimeException") {
      console.log(
        '`${caught.name}` - "Error occurred while sending Converse request.',
      );
      throw caught;
    }
  }
}

// Processes the response received via Amazon Bedrock and performs the necessary actions based on the stop reason.
// param "response" - The model's response returned via Amazon Bedrock.
// param "messages" - The conversation history.
// param "max_recursions" - The maximum number of recursive calls allowed.
async function ProcessModelResponseAsync(response, messages, max_recursions) {
  if (max_recursions <= 0) {
    await HandleToolUseAsync(response, messages);
  }
  if (response.stopReason === "tool_use") {
    await HandleToolUseAsync(response, messages, max_recursions - 1);
  }
  if (response.stopReason === "end_turn") {
    const messageToPrint = response.output.message.content[0].text;
    console.log(messageToPrint.replace(/<[^>]+>/g, ""));
  }
}
// Handles the tool use case by invoking the specified tool and sending the tool's response back to Bedrock.
// The tool response is appended to the conversation, and the conversation is sent back to Amazon Bedrock for further processing.
// param "response" - the model's response containing the tool use request.
// param "messages" - the conversation history.
// param "max_recursions" - The maximum number of recursive calls allowed.
async function HandleToolUseAsync(response, messages, max_recursions) {
  const toolResultFinal = [];
  try {
    const output_message = response.output.message;
    messages.push(output_message);
    const toolRequests = output_message.content;
    const toolMessage = toolRequests[0].text;
    console.log(toolMessage.replace(/<[^>]+>/g, ""));
    for (const toolRequest of toolRequests) {
      if (Object.hasOwn(toolRequest, "toolUse")) {
        const toolUse = toolRequest.toolUse;
        const latitude = toolUse.input.latitude;
        const longitude = toolUse.input.longitude;
        const toolUseID = toolUse.toolUseId;
        console.log(
          `Requesting tool ${toolUse.name}, Tool use id ${toolUseID}`,
        );
        if (toolUse.name === "Weather_Tool") {
          try {
            const current_weather = await callWeatherTool(
              longitude,
              latitude,
            ).then((current_weather) => current_weather);
            const currentWeather = current_weather;
            const toolResult = {
              toolResult: {
                toolUseId: toolUseID,
                content: [{ json: currentWeather }],
              },
            };
            toolResultFinal.push(toolResult);
          } catch (err) {
            console.log("An error occurred. ", err);
          }
        }
      }
    }

    const toolResultMessage = {
      role: "user",
      content: toolResultFinal,
    };
    messages.push(toolResultMessage);
    // Send the conversation to Amazon Bedrock
    await ProcessModelResponseAsync(
      await SendConversationtoBedrock(messages),
      messages,
    );
  } catch (error) {
    console.log("An error occurred. ", error);
  }
}
// Call the Weathertool.
// param = longitude of location
// param = latitude of location
async function callWeatherTool(longitude, latitude) {
  // Open-Meteo API endpoint
  const apiUrl = `https://api.open-meteo.com/v1/forecast?latitude=${latitude}&longitude=${longitude}&current_weather=true`;

  // Fetch the weather data.
  return fetch(apiUrl)
    .then((response) => {
      return response.json().then((current_weather) => {
        return current_weather;
      });
    })
    .catch((error) => {
      console.error("Error fetching weather data:", error);
    });
}
/**
 * Used repeatedly to have the user press enter.
 * @type {ScenarioInput}
 */
const pressEnter = new ScenarioInput("continue", "Press Enter to continue", {
  type: "input",
  default: "",
});

const greet = new ScenarioOutput(
  "greet",
  "Welcome to the Amazon Bedrock Tool Use demo! \n" +
    "This assistant provides current weather information for user-specified locations. " +
    "You can ask for weather details by providing the location name or coordinates." +
    "Weather information will be provided using a custom Tool and open-meteo API." +
    "For the purposes of this example, we'll use in order the questions in ./questions.json :\n" +
    "What's the weather like in Seattle? " +
    "What's the best kind of cat? " +
    "Where is the warmest city in Washington State right now? " +
    "What's the warmest city in California right now?\n" +
    "To exit the program, simply type 'x' and press Enter.\n" +
    "Have fun and experiment with the app by editing the questions in ./questions.json! " +
    "P.S.: You're not limited to single locations, or even to using English! ",

  { header: true },
);
const displayAskQuestion1 = new ScenarioOutput(
  "displayAskQuestion1",
  "Press enter to ask question number 1 (default is 'What's the weather like in Seattle?')",
);

const askQuestion1 = new ScenarioAction(
  "askQuestion1",
  async (/** @type {State} */ state) => {
    const userMessage1 = data.questions["question-1"];
    await askQuestion(userMessage1);
  },
);

const displayAskQuestion2 = new ScenarioOutput(
  "displayAskQuestion2",
  "Press enter to ask question number 2 (default is 'What's the best kind of cat?')",
);

const askQuestion2 = new ScenarioAction(
  "askQuestion2",
  async (/** @type {State} */ state) => {
    const userMessage2 = data.questions["question-2"];
    await askQuestion(userMessage2);
  },
);
const displayAskQuestion3 = new ScenarioOutput(
  "displayAskQuestion3",
  "Press enter to ask question number 3 (default is 'Where is the warmest city in Washington State right now?')",
);

const askQuestion3 = new ScenarioAction(
  "askQuestion3",
  async (/** @type {State} */ state) => {
    const userMessage3 = data.questions["question-3"];
    await askQuestion(userMessage3);
  },
);

const displayAskQuestion4 = new ScenarioOutput(
  "displayAskQuestion4",
  "Press enter to ask question number 4 (default is 'What's the warmest city in California right now?')",
);

const askQuestion4 = new ScenarioAction(
  "askQuestion4",
  async (/** @type {State} */ state) => {
    const userMessage4 = data.questions["question-4"];
    await askQuestion(userMessage4);
  },
);

const goodbye = new ScenarioOutput(
  "goodbye",
  "Thank you for checking out the Amazon Bedrock Tool Use demo. We hope you\n" +
    "learned something new, or got some inspiration for your own apps today!\n" +
    "For more Bedrock examples in different programming languages, have a look at:\n" +
    "https://docs.aws.amazon.com/bedrock/latest/userguide/service_code_examples.html",
);

const myScenario = new Scenario("Converse Tool Scenario", [
  greet,
  pressEnter,
  displayAskQuestion1,
  askQuestion1,
  pressEnter,
  displayAskQuestion2,
  askQuestion2,
  pressEnter,
  displayAskQuestion3,
  askQuestion3,
  pressEnter,
  displayAskQuestion4,
  askQuestion4,
  pressEnter,
  goodbye,
]);

/** @type {{ stepHandlerOptions: StepHandlerOptions }} */
export const main = async (stepHandlerOptions) => {
  await myScenario.run(stepHandlerOptions);
};

// Invoke main function if this file was run directly.
if (process.argv[1] === fileURLToPath(import.meta.url)) {
  const { values } = parseArgs({
    options: {
      yes: {
        type: "boolean",
        short: "y",
      },
    },
  });
  main({ confirmAll: values.yes });
}
```
+  如需 API 詳細資訊，請參閱《適用於 JavaScript 的 AWS SDK API 參考》**中的 [Converse](https://docs.aws.amazon.com/AWSJavaScriptSDK/v3/latest/client/bedrock-runtime/command/ConverseCommand)。

------
#### [ Python ]

**適用於 Python 的 SDK (Boto3)**  
 GitHub 上提供更多範例。尋找完整範例，並了解如何在 [AWS 程式碼範例儲存庫](https://github.com/awsdocs/aws-doc-sdk-examples/tree/main/python/example_code/bedrock-runtime#code-examples)中設定和執行。
示範的主要執行指令碼。此指令碼會協調使用者、Amazon Bedrock Converse API 和天氣工具之間的對話。  

```
"""
This demo illustrates a tool use scenario using Amazon Bedrock's Converse API and a weather tool.
The script interacts with a foundation model on Amazon Bedrock to provide weather information based on user
input. It uses the Open-Meteo API (https://open-meteo.com) to retrieve current weather data for a given location.
"""

import boto3
import logging
from enum import Enum

import utils.tool_use_print_utils as output
import weather_tool

logging.basicConfig(level=logging.INFO, format="%(message)s")

AWS_REGION = "us-east-1"


# For the most recent list of models supported by the Converse API's tool use functionality, visit:
# https://docs.aws.amazon.com/bedrock/latest/userguide/conversation-inference.html
class SupportedModels(Enum):
    CLAUDE_OPUS = "anthropic.claude-3-opus-20240229-v1:0"
    CLAUDE_SONNET = "anthropic.claude-3-sonnet-20240229-v1:0"
    CLAUDE_HAIKU = "anthropic.claude-3-haiku-20240307-v1:0"
    COHERE_COMMAND_R = "cohere.command-r-v1:0"
    COHERE_COMMAND_R_PLUS = "cohere.command-r-plus-v1:0"


# Set the model ID, e.g., Claude 3 Haiku.
MODEL_ID = SupportedModels.CLAUDE_HAIKU.value

SYSTEM_PROMPT = """
You are a weather assistant that provides current weather data for user-specified locations using only
the Weather_Tool, which expects latitude and longitude. Infer the coordinates from the location yourself.
If the user provides coordinates, infer the approximate location and refer to it in your response.
To use the tool, you strictly apply the provided tool specification.

- Explain your step-by-step process, and give brief updates before each step.
- Only use the Weather_Tool for data. Never guess or make up information. 
- Repeat the tool use for subsequent requests if necessary.
- If the tool errors, apologize, explain weather is unavailable, and suggest other options.
- Report temperatures in °C (°F) and wind in km/h (mph). Keep weather reports concise. Sparingly use
  emojis where appropriate.
- Only respond to weather queries. Remind off-topic users of your purpose. 
- Never claim to search online, access external data, or use tools besides Weather_Tool.
- Complete the entire process until you have all required data before sending the complete response.
"""

# The maximum number of recursive calls allowed in the tool_use_demo function.
# This helps prevent infinite loops and potential performance issues.
MAX_RECURSIONS = 5


class ToolUseDemo:
    """
    Demonstrates the tool use feature with the Amazon Bedrock Converse API.
    """

    def __init__(self):
        # Prepare the system prompt
        self.system_prompt = [{"text": SYSTEM_PROMPT}]

        # Prepare the tool configuration with the weather tool's specification
        self.tool_config = {"tools": [weather_tool.get_tool_spec()]}

        # Create a Bedrock Runtime client in the specified AWS Region.
        self.bedrockRuntimeClient = boto3.client(
            "bedrock-runtime", region_name=AWS_REGION
        )

    def run(self):
        """
        Starts the conversation with the user and handles the interaction with Bedrock.
        """
        # Print the greeting and a short user guide
        output.header()

        # Start with an emtpy conversation
        conversation = []

        # Get the first user input
        user_input = self._get_user_input()

        while user_input is not None:
            # Create a new message with the user input and append it to the conversation
            message = {"role": "user", "content": [{"text": user_input}]}
            conversation.append(message)

            # Send the conversation to Amazon Bedrock
            bedrock_response = self._send_conversation_to_bedrock(conversation)

            # Recursively handle the model's response until the model has returned
            # its final response or the recursion counter has reached 0
            self._process_model_response(
                bedrock_response, conversation, max_recursion=MAX_RECURSIONS
            )

            # Repeat the loop until the user decides to exit the application
            user_input = self._get_user_input()

        output.footer()

    def _send_conversation_to_bedrock(self, conversation):
        """
        Sends the conversation, the system prompt, and the tool spec to Amazon Bedrock, and returns the response.

        :param conversation: The conversation history including the next message to send.
        :return: The response from Amazon Bedrock.
        """
        output.call_to_bedrock(conversation)

        # Send the conversation, system prompt, and tool configuration, and return the response
        return self.bedrockRuntimeClient.converse(
            modelId=MODEL_ID,
            messages=conversation,
            system=self.system_prompt,
            toolConfig=self.tool_config,
        )

    def _process_model_response(
        self, model_response, conversation, max_recursion=MAX_RECURSIONS
    ):
        """
        Processes the response received via Amazon Bedrock and performs the necessary actions
        based on the stop reason.

        :param model_response: The model's response returned via Amazon Bedrock.
        :param conversation: The conversation history.
        :param max_recursion: The maximum number of recursive calls allowed.
        """

        if max_recursion <= 0:
            # Stop the process, the number of recursive calls could indicate an infinite loop
            logging.warning(
                "Warning: Maximum number of recursions reached. Please try again."
            )
            exit(1)

        # Append the model's response to the ongoing conversation
        message = model_response["output"]["message"]
        conversation.append(message)

        if model_response["stopReason"] == "tool_use":
            # If the stop reason is "tool_use", forward everything to the tool use handler
            self._handle_tool_use(message, conversation, max_recursion)

        if model_response["stopReason"] == "end_turn":
            # If the stop reason is "end_turn", print the model's response text, and finish the process
            output.model_response(message["content"][0]["text"])
            return

    def _handle_tool_use(
        self, model_response, conversation, max_recursion=MAX_RECURSIONS
    ):
        """
        Handles the tool use case by invoking the specified tool and sending the tool's response back to Bedrock.
        The tool response is appended to the conversation, and the conversation is sent back to Amazon Bedrock for further processing.

        :param model_response: The model's response containing the tool use request.
        :param conversation: The conversation history.
        :param max_recursion: The maximum number of recursive calls allowed.
        """

        # Initialize an empty list of tool results
        tool_results = []

        # The model's response can consist of multiple content blocks
        for content_block in model_response["content"]:
            if "text" in content_block:
                # If the content block contains text, print it to the console
                output.model_response(content_block["text"])

            if "toolUse" in content_block:
                # If the content block is a tool use request, forward it to the tool
                tool_response = self._invoke_tool(content_block["toolUse"])

                # Add the tool use ID and the tool's response to the list of results
                tool_results.append(
                    {
                        "toolResult": {
                            "toolUseId": (tool_response["toolUseId"]),
                            "content": [{"json": tool_response["content"]}],
                        }
                    }
                )

        # Embed the tool results in a new user message
        message = {"role": "user", "content": tool_results}

        # Append the new message to the ongoing conversation
        conversation.append(message)

        # Send the conversation to Amazon Bedrock
        response = self._send_conversation_to_bedrock(conversation)

        # Recursively handle the model's response until the model has returned
        # its final response or the recursion counter has reached 0
        self._process_model_response(response, conversation, max_recursion - 1)

    def _invoke_tool(self, payload):
        """
        Invokes the specified tool with the given payload and returns the tool's response.
        If the requested tool does not exist, an error message is returned.

        :param payload: The payload containing the tool name and input data.
        :return: The tool's response or an error message.
        """
        tool_name = payload["name"]

        if tool_name == "Weather_Tool":
            input_data = payload["input"]
            output.tool_use(tool_name, input_data)

            # Invoke the weather tool with the input data provided by
            response = weather_tool.fetch_weather_data(input_data)
        else:
            error_message = (
                f"The requested tool with name '{tool_name}' does not exist."
            )
            response = {"error": "true", "message": error_message}

        return {"toolUseId": payload["toolUseId"], "content": response}

    @staticmethod
    def _get_user_input(prompt="Your weather info request"):
        """
        Prompts the user for input and returns the user's response.
        Returns None if the user enters 'x' to exit.

        :param prompt: The prompt to display to the user.
        :return: The user's input or None if the user chooses to exit.
        """
        output.separator()
        user_input = input(f"{prompt} (x to exit): ")

        if user_input == "":
            prompt = "Please enter your weather info request, e.g. the name of a city"
            return ToolUseDemo._get_user_input(prompt)

        elif user_input.lower() == "x":
            return None

        else:
            return user_input


if __name__ == "__main__":
    tool_use_demo = ToolUseDemo()
    tool_use_demo.run()
```
示範時使用的天氣工具。此指令碼定義工具規格，並實作邏輯，以從 Open-Meteo API 用來擷取天氣資料。  

```
import requests
from requests.exceptions import RequestException


def get_tool_spec():
    """
    Returns the JSON Schema specification for the Weather tool. The tool specification
    defines the input schema and describes the tool's functionality.
    For more information, see https://json-schema.org/understanding-json-schema/reference.

    :return: The tool specification for the Weather tool.
    """
    return {
        "toolSpec": {
            "name": "Weather_Tool",
            "description": "Get the current weather for a given location, based on its WGS84 coordinates.",
            "inputSchema": {
                "json": {
                    "type": "object",
                    "properties": {
                        "latitude": {
                            "type": "string",
                            "description": "Geographical WGS84 latitude of the location.",
                        },
                        "longitude": {
                            "type": "string",
                            "description": "Geographical WGS84 longitude of the location.",
                        },
                    },
                    "required": ["latitude", "longitude"],
                }
            },
        }
    }


def fetch_weather_data(input_data):
    """
    Fetches weather data for the given latitude and longitude using the Open-Meteo API.
    Returns the weather data or an error message if the request fails.

    :param input_data: The input data containing the latitude and longitude.
    :return: The weather data or an error message.
    """
    endpoint = "https://api.open-meteo.com/v1/forecast"
    latitude = input_data.get("latitude")
    longitude = input_data.get("longitude", "")
    params = {"latitude": latitude, "longitude": longitude, "current_weather": True}

    try:
        response = requests.get(endpoint, params=params)
        weather_data = {"weather_data": response.json()}
        response.raise_for_status()
        return weather_data
    except RequestException as e:
        return e.response.json()
    except Exception as e:
        return {"error": type(e), "message": str(e)}
```
+  如需 API 詳細資訊，請參閱《AWS SDK for Python (Boto3) API 參考》**中的 [Converse](https://docs.aws.amazon.com/goto/boto3/bedrock-runtime-2023-09-30/Converse)。

------
#### [ Rust ]

**適用於 Rust 的 SDK**  
 GitHub 上提供更多範例。尋找完整範例，並了解如何在 [AWS 程式碼範例儲存庫](https://github.com/awsdocs/aws-doc-sdk-examples/tree/main/rustv1/examples/bedrock-runtime#code-examples)中設定和執行。
示範的主要案例和邏輯。這會協調使用者、Amazon Bedrock Converse API 和天氣工具之間的對話。  

```
#[derive(Debug)]
#[allow(dead_code)]
struct InvokeToolResult(String, ToolResultBlock);
struct ToolUseScenario {
    client: Client,
    conversation: Vec<Message>,
    system_prompt: SystemContentBlock,
    tool_config: ToolConfiguration,
}

impl ToolUseScenario {
    fn new(client: Client) -> Self {
        let system_prompt = SystemContentBlock::Text(SYSTEM_PROMPT.into());
        let tool_config = ToolConfiguration::builder()
            .tools(Tool::ToolSpec(
                ToolSpecification::builder()
                    .name(TOOL_NAME)
                    .description(TOOL_DESCRIPTION)
                    .input_schema(ToolInputSchema::Json(make_tool_schema()))
                    .build()
                    .unwrap(),
            ))
            .build()
            .unwrap();

        ToolUseScenario {
            client,
            conversation: vec![],
            system_prompt,
            tool_config,
        }
    }

    async fn run(&mut self) -> Result<(), ToolUseScenarioError> {
        loop {
            let input = get_input().await?;
            if input.is_none() {
                break;
            }

            let message = Message::builder()
                .role(User)
                .content(ContentBlock::Text(input.unwrap()))
                .build()
                .map_err(ToolUseScenarioError::from)?;
            self.conversation.push(message);

            let response = self.send_to_bedrock().await?;

            self.process_model_response(response).await?;
        }

        Ok(())
    }

    async fn send_to_bedrock(&mut self) -> Result<ConverseOutput, ToolUseScenarioError> {
        debug!("Sending conversation to bedrock");
        self.client
            .converse()
            .model_id(MODEL_ID)
            .set_messages(Some(self.conversation.clone()))
            .system(self.system_prompt.clone())
            .tool_config(self.tool_config.clone())
            .send()
            .await
            .map_err(ToolUseScenarioError::from)
    }

    async fn process_model_response(
        &mut self,
        mut response: ConverseOutput,
    ) -> Result<(), ToolUseScenarioError> {
        let mut iteration = 0;

        while iteration < MAX_RECURSIONS {
            iteration += 1;
            let message = if let Some(ref output) = response.output {
                if output.is_message() {
                    Ok(output.as_message().unwrap().clone())
                } else {
                    Err(ToolUseScenarioError(
                        "Converse Output is not a message".into(),
                    ))
                }
            } else {
                Err(ToolUseScenarioError("Missing Converse Output".into()))
            }?;

            self.conversation.push(message.clone());

            match response.stop_reason {
                StopReason::ToolUse => {
                    response = self.handle_tool_use(&message).await?;
                }
                StopReason::EndTurn => {
                    print_model_response(&message.content[0])?;
                    return Ok(());
                }
                _ => (),
            }
        }

        Err(ToolUseScenarioError(
            "Exceeded MAX_ITERATIONS when calling tools".into(),
        ))
    }

    async fn handle_tool_use(
        &mut self,
        message: &Message,
    ) -> Result<ConverseOutput, ToolUseScenarioError> {
        let mut tool_results: Vec<ContentBlock> = vec![];

        for block in &message.content {
            match block {
                ContentBlock::Text(_) => print_model_response(block)?,
                ContentBlock::ToolUse(tool) => {
                    let tool_response = self.invoke_tool(tool).await?;
                    tool_results.push(ContentBlock::ToolResult(tool_response.1));
                }
                _ => (),
            };
        }

        let message = Message::builder()
            .role(User)
            .set_content(Some(tool_results))
            .build()?;
        self.conversation.push(message);

        self.send_to_bedrock().await
    }

    async fn invoke_tool(
        &mut self,
        tool: &ToolUseBlock,
    ) -> Result<InvokeToolResult, ToolUseScenarioError> {
        match tool.name() {
            TOOL_NAME => {
                println!(
                    "\x1b[0;90mExecuting tool: {TOOL_NAME} with input: {:?}...\x1b[0m",
                    tool.input()
                );
                let content = fetch_weather_data(tool).await?;
                println!(
                    "\x1b[0;90mTool responded with {:?}\x1b[0m",
                    content.content()
                );
                Ok(InvokeToolResult(tool.tool_use_id.clone(), content))
            }
            _ => Err(ToolUseScenarioError(format!(
                "The requested tool with name {} does not exist",
                tool.name()
            ))),
        }
    }
}

#[tokio::main]
async fn main() {
    tracing_subscriber::fmt::init();
    let sdk_config = aws_config::defaults(BehaviorVersion::latest())
        .region(CLAUDE_REGION)
        .load()
        .await;
    let client = Client::new(&sdk_config);

    let mut scenario = ToolUseScenario::new(client);

    header();
    if let Err(err) = scenario.run().await {
        println!("There was an error running the scenario! {}", err.0)
    }
    footer();
}
```
示範時使用的天氣工具。此指令碼定義工具規格，並實作邏輯，以從 Open-Meteo API 用來擷取天氣資料。  

```
const ENDPOINT: &str = "https://api.open-meteo.com/v1/forecast";
async fn fetch_weather_data(
    tool_use: &ToolUseBlock,
) -> Result<ToolResultBlock, ToolUseScenarioError> {
    let input = tool_use.input();
    let latitude = input
        .as_object()
        .unwrap()
        .get("latitude")
        .unwrap()
        .as_string()
        .unwrap();
    let longitude = input
        .as_object()
        .unwrap()
        .get("longitude")
        .unwrap()
        .as_string()
        .unwrap();
    let params = [
        ("latitude", latitude),
        ("longitude", longitude),
        ("current_weather", "true"),
    ];

    debug!("Calling {ENDPOINT} with {params:?}");

    let response = reqwest::Client::new()
        .get(ENDPOINT)
        .query(&params)
        .send()
        .await
        .map_err(|e| ToolUseScenarioError(format!("Error requesting weather: {e:?}")))?
        .error_for_status()
        .map_err(|e| ToolUseScenarioError(format!("Failed to request weather: {e:?}")))?;

    debug!("Response: {response:?}");

    let bytes = response
        .bytes()
        .await
        .map_err(|e| ToolUseScenarioError(format!("Error reading response: {e:?}")))?;

    let result = String::from_utf8(bytes.to_vec())
        .map_err(|_| ToolUseScenarioError("Response was not utf8".into()))?;

    Ok(ToolResultBlock::builder()
        .tool_use_id(tool_use.tool_use_id())
        .content(ToolResultContentBlock::Text(result))
        .build()?)
}
```
可列印訊息內容區塊的公用程式。  

```
fn print_model_response(block: &ContentBlock) -> Result<(), ToolUseScenarioError> {
    if block.is_text() {
        let text = block.as_text().unwrap();
        println!("\x1b[0;90mThe model's response:\x1b[0m\n{text}");
        Ok(())
    } else {
        Err(ToolUseScenarioError(format!(
            "Content block is not text ({block:?})"
        )))
    }
}
```
使用陳述式、錯誤公用程式和常數。  

```
use std::{collections::HashMap, io::stdin};

use aws_config::BehaviorVersion;
use aws_sdk_bedrockruntime::{
    error::{BuildError, SdkError},
    operation::converse::{ConverseError, ConverseOutput},
    types::{
        ContentBlock, ConversationRole::User, Message, StopReason, SystemContentBlock, Tool,
        ToolConfiguration, ToolInputSchema, ToolResultBlock, ToolResultContentBlock,
        ToolSpecification, ToolUseBlock,
    },
    Client,
};
use aws_smithy_runtime_api::http::Response;
use aws_smithy_types::Document;
use tracing::debug;

// Set the model ID, e.g., Claude 3 Haiku.
const MODEL_ID: &str = "anthropic.claude-3-haiku-20240307-v1:0";
const CLAUDE_REGION: &str = "us-east-1";

const SYSTEM_PROMPT: &str = "You are a weather assistant that provides current weather data for user-specified locations using only
the Weather_Tool, which expects latitude and longitude. Infer the coordinates from the location yourself.
If the user provides coordinates, infer the approximate location and refer to it in your response.
To use the tool, you strictly apply the provided tool specification.

- Explain your step-by-step process, and give brief updates before each step.
- Only use the Weather_Tool for data. Never guess or make up information. 
- Repeat the tool use for subsequent requests if necessary.
- If the tool errors, apologize, explain weather is unavailable, and suggest other options.
- Report temperatures in °C (°F) and wind in km/h (mph). Keep weather reports concise. Sparingly use
  emojis where appropriate.
- Only respond to weather queries. Remind off-topic users of your purpose. 
- Never claim to search online, access external data, or use tools besides Weather_Tool.
- Complete the entire process until you have all required data before sending the complete response.
";

// The maximum number of recursive calls allowed in the tool_use_demo function.
// This helps prevent infinite loops and potential performance issues.
const MAX_RECURSIONS: i8 = 5;

const TOOL_NAME: &str = "Weather_Tool";
const TOOL_DESCRIPTION: &str =
    "Get the current weather for a given location, based on its WGS84 coordinates.";
fn make_tool_schema() -> Document {
    Document::Object(HashMap::<String, Document>::from([
        ("type".into(), Document::String("object".into())),
        (
            "properties".into(),
            Document::Object(HashMap::from([
                (
                    "latitude".into(),
                    Document::Object(HashMap::from([
                        ("type".into(), Document::String("string".into())),
                        (
                            "description".into(),
                            Document::String("Geographical WGS84 latitude of the location.".into()),
                        ),
                    ])),
                ),
                (
                    "longitude".into(),
                    Document::Object(HashMap::from([
                        ("type".into(), Document::String("string".into())),
                        (
                            "description".into(),
                            Document::String(
                                "Geographical WGS84 longitude of the location.".into(),
                            ),
                        ),
                    ])),
                ),
            ])),
        ),
        (
            "required".into(),
            Document::Array(vec![
                Document::String("latitude".into()),
                Document::String("longitude".into()),
            ]),
        ),
    ]))
}

#[derive(Debug)]
struct ToolUseScenarioError(String);
impl std::fmt::Display for ToolUseScenarioError {
    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
        write!(f, "Tool use error with '{}'. Reason: {}", MODEL_ID, self.0)
    }
}
impl From<&str> for ToolUseScenarioError {
    fn from(value: &str) -> Self {
        ToolUseScenarioError(value.into())
    }
}
impl From<BuildError> for ToolUseScenarioError {
    fn from(value: BuildError) -> Self {
        ToolUseScenarioError(value.to_string().clone())
    }
}
impl From<SdkError<ConverseError, Response>> for ToolUseScenarioError {
    fn from(value: SdkError<ConverseError, Response>) -> Self {
        ToolUseScenarioError(match value.as_service_error() {
            Some(value) => value.meta().message().unwrap_or("Unknown").into(),
            None => "Unknown".into(),
        })
    }
}
```
+  如需 API 詳細資訊，請參閱《AWS SDK for Rust API 參考》**中的 [Converse](https://docs.rs/aws-sdk-bedrockruntime/latest/aws_sdk_bedrockruntime/client/struct.Client.html#method.converse)。

------

如需 AWS SDK 開發人員指南和程式碼範例的完整清單，請參閱 [搭配 AWS SDK 使用 Amazon Bedrock](sdk-general-information-section.md)。此主題也包含有關入門的資訊和舊版 SDK 的詳細資訊。