Codebeispiele für Amazon Bedrock Runtime mit AWS SDKs

Die folgenden Codebeispiele zeigen, wie Amazon Bedrock Runtime mit einem AWS Software Development Kit (SDK) verwendet wird.

Szenarien sind Code-Beispiele, die Ihnen zeigen, wie Sie bestimmte Aufgaben ausführen, indem Sie mehrere Funktionen innerhalb eines Services aufrufen oder mit anderen AWS-Services kombinieren.

Eine vollständige Liste der AWS SDK-Entwicklerhandbücher und Codebeispiele finden Sie unterAmazon Bedrock mit einem AWS SDK verwenden. Dieses Thema enthält auch Informationen zu den ersten Schritten und Details zu früheren SDK-Versionen.

Erste Schritte

Hallo Amazon Bedrock

Die folgenden Codebeispiele zeigen, wie Sie mit Amazon Bedrock beginnen können.

Go

SDK für Go V2

Anmerkung

Es gibt noch mehr dazu. GitHub Hier finden Sie das vollständige Beispiel und erfahren, wie Sie das AWS -Code-Beispiel- einrichten und ausführen.

package main

import (
	"context"
	"encoding/json"
	"flag"
	"fmt"
	"log"
	"os"
	"strings"

	"github.com/aws/aws-sdk-go-v2/aws"
	"github.com/aws/aws-sdk-go-v2/config"
	"github.com/aws/aws-sdk-go-v2/service/bedrockruntime"
)

// Each model provider defines their own individual request and response formats.
// For the format, ranges, and default values for the different models, refer to:
// https://docs.aws.amazon.com/bedrock/latest/userguide/model-parameters.html

type ClaudeRequest struct {
	Prompt            string `json:"prompt"`
	MaxTokensToSample int    `json:"max_tokens_to_sample"`
	// Omitting optional request parameters
}

type ClaudeResponse struct {
	Completion string `json:"completion"`
}

// main uses the AWS SDK for Go (v2) to create an Amazon Bedrock Runtime client
// and invokes Anthropic Claude 2 inside your account and the chosen region.
// This example uses the default settings specified in your shared credentials
// and config files.
func main() {

	region := flag.String("region", "us-east-1", "The AWS region")
	flag.Parse()

	fmt.Printf("Using AWS region: %s\n", *region)

	ctx := context.Background()
	sdkConfig, err := config.LoadDefaultConfig(ctx, config.WithRegion(*region))
	if err != nil {
		fmt.Println("Couldn't load default configuration. Have you set up your AWS account?")
		fmt.Println(err)
		return
	}

	client := bedrockruntime.NewFromConfig(sdkConfig)

	modelId := "anthropic.claude-v2"

	prompt := "Hello, how are you today?"

	// Anthropic Claude requires you to enclose the prompt as follows:
	prefix := "Human: "
	postfix := "\n\nAssistant:"
	wrappedPrompt := prefix + prompt + postfix

	request := ClaudeRequest{
		Prompt:            wrappedPrompt,
		MaxTokensToSample: 200,
	}

	body, err := json.Marshal(request)
	if err != nil {
		log.Panicln("Couldn't marshal the request: ", err)
	}

	result, err := client.InvokeModel(ctx, &bedrockruntime.InvokeModelInput{
		ModelId:     aws.String(modelId),
		ContentType: aws.String("application/json"),
		Body:        body,
	})

	if err != nil {
		errMsg := err.Error()
		if strings.Contains(errMsg, "no such host") {
			fmt.Printf("Error: The Bedrock service is not available in the selected region. Please double-check the service availability for your region at https://aws.amazon.com/about-aws/global-infrastructure/regional-product-services/.\n")
		} else if strings.Contains(errMsg, "Could not resolve the foundation model") {
			fmt.Printf("Error: Could not resolve the foundation model from model identifier: \"%v\". Please verify that the requested model exists and is accessible within the specified region.\n", modelId)
		} else {
			fmt.Printf("Error: Couldn't invoke Anthropic Claude. Here's why: %v\n", err)
		}
		os.Exit(1)
	}

	var response ClaudeResponse

	err = json.Unmarshal(result.Body, &response)

	if err != nil {
		log.Fatal("failed to unmarshal", err)
	}
	fmt.Println("Prompt:\n", prompt)
	fmt.Println("Response from Anthropic Claude:\n", response.Completion)
}

• Einzelheiten zur API finden Sie InvokeModelin der AWS SDK für Go API-Referenz.

JavaScript

SDK für JavaScript (v3)

Anmerkung

Es gibt noch mehr dazu GitHub. Hier finden Sie das vollständige Beispiel und erfahren, wie Sie das AWS -Code-Beispiel- einrichten und ausführen.

/**
 * @typedef {Object} Content
 * @property {string} text
 *
 * @typedef {Object} Usage
 * @property {number} input_tokens
 * @property {number} output_tokens
 *
 * @typedef {Object} ResponseBody
 * @property {Content[]} content
 * @property {Usage} usage
 */

import { fileURLToPath } from "node:url";
import {
  BedrockRuntimeClient,
  InvokeModelCommand,
} from "@aws-sdk/client-bedrock-runtime";

const AWS_REGION = "us-east-1";

const MODEL_ID = "anthropic.claude-3-haiku-20240307-v1:0";
const PROMPT = "Hi. In a short paragraph, explain what you can do.";

const hello = async () => {
  console.log("=".repeat(35));
  console.log("Welcome to the Amazon Bedrock demo!");
  console.log("=".repeat(35));

  console.log("Model: Anthropic Claude 3 Haiku");
  console.log(`Prompt: ${PROMPT}\n`);
  console.log("Invoking model...\n");

  // Create a new Bedrock Runtime client instance.
  const client = new BedrockRuntimeClient({ region: AWS_REGION });

  // Prepare the payload for the model.
  const payload = {
    anthropic_version: "bedrock-2023-05-31",
    max_tokens: 1000,
    messages: [{ role: "user", content: [{ type: "text", text: PROMPT }] }],
  };

  // Invoke Claude with the payload and wait for the response.
  const apiResponse = await client.send(
    new InvokeModelCommand({
      contentType: "application/json",
      body: JSON.stringify(payload),
      modelId: MODEL_ID,
    }),
  );

  // Decode and return the response(s)
  const decodedResponseBody = new TextDecoder().decode(apiResponse.body);
  /** @type {ResponseBody} */
  const responseBody = JSON.parse(decodedResponseBody);
  const responses = responseBody.content;

  if (responses.length === 1) {
    console.log(`Response: ${responses[0].text}`);
  } else {
    console.log("Haiku returned multiple responses:");
    console.log(responses);
  }

  console.log(`\nNumber of input tokens:   ${responseBody.usage.input_tokens}`);
  console.log(`Number of output tokens: ${responseBody.usage.output_tokens}`);
};

if (process.argv[1] === fileURLToPath(import.meta.url)) {
  await hello();
}

• Einzelheiten zur API finden Sie InvokeModelin der AWS SDK for JavaScript API-Referenz.

Python

SDK für Python (Boto3)

Anmerkung

Es gibt noch mehr dazu GitHub. Hier finden Sie das vollständige Beispiel und erfahren, wie Sie das AWS -Code-Beispiel- einrichten und ausführen.

Senden Sie mit der InvokeModel Operation eine Aufforderung an ein Modell.

"""
Uses the Amazon Bedrock runtime client InvokeModel operation to send a prompt to a model.
"""
import logging
import json
import boto3


from botocore.exceptions import ClientError


logging.basicConfig(level=logging.INFO)
logger = logging.getLogger(__name__)


def invoke_model(brt, model_id, prompt):
    """
    Invokes the specified model with the supplied prompt.
    param brt: A bedrock runtime boto3 client
    param model_id: The model ID for the model that you want to use.
    param prompt: The prompt that you want to send to the model.

    :return: The text response from the model.
    """

    # Format the request payload using the model's native structure.
    native_request = {
        "inputText": prompt,
        "textGenerationConfig": {
            "maxTokenCount": 512,
            "temperature": 0.5,
            "topP": 0.9
        }
    }

    # Convert the native request to JSON.
    request = json.dumps(native_request)

    try:
        # Invoke the model with the request.
        response = brt.invoke_model(modelId=model_id, body=request)

        # Decode the response body.
        model_response = json.loads(response["body"].read())

        # Extract and print the response text.
        response_text = model_response["results"][0]["outputText"]
        return response_text

    except (ClientError, Exception) as e:
        print(f"ERROR: Can't invoke '{model_id}'. Reason: {e}")
        raise


def main():
    """Entry point for the example. Uses the AWS SDK for Python (Boto3)
    to create an Amazon Bedrock runtime client. Then sends a prompt to a model
    in the region set in the callers profile and credentials.
    """

    # Create an Amazon Bedrock Runtime client.
    brt = boto3.client("bedrock-runtime")

    # Set the model ID, e.g., Amazon Titan Text G1 - Express.
    model_id = "amazon.titan-text-express-v1"

    # Define the prompt for the model.
    prompt = "Describe the purpose of a 'hello world' program in one line."

    # Send the prompt to the model.
    response = invoke_model(brt, model_id, prompt)

    print(f"Response: {response}")

    logger.info("Done.")


if __name__ == "__main__":
    main()

Senden Sie mit der Converse-Operation eine Benutzernachricht an ein Modell.

"""
Uses the Amazon Bedrock runtime client Converse operation to send a user message to a model.
"""
import logging
import boto3

from botocore.exceptions import ClientError


logging.basicConfig(level=logging.INFO)
logger = logging.getLogger(__name__)


def converse(brt, model_id, user_message):
    """
    Uses the Converse operation to send a user message to the supplied model.
    param brt: A bedrock runtime boto3 client
    param model_id: The model ID for the model that you want to use.
    param user message: The user message that you want to send to the model.

    :return: The text response from the model.
    """

    # Format the request payload using the model's native structure.
    conversation = [
    {
        "role": "user",
        "content": [{"text": user_message}],
    }
]

    try:
        # Send the message to the model, using a basic inference configuration.
        response = brt.converse(
            modelId=model_id,
            messages=conversation,
            inferenceConfig={"maxTokens": 512, "temperature": 0.5, "topP": 0.9},
        )

        # Extract and print the response text.
        response_text = response["output"]["message"]["content"][0]["text"]
        return response_text

    except (ClientError, Exception) as e:
        print(f"ERROR: Can't invoke '{model_id}'. Reason: {e}")
        raise


def main():
    """Entry point for the example. Uses the AWS SDK for Python (Boto3)
    to create an Amazon Bedrock runtime client. Then sends a user message to a model
    in the region set in the callers profile and credentials.
    """

    # Create an Amazon Bedrock Runtime client.
    brt = boto3.client("bedrock-runtime")

    # Set the model ID, e.g., Amazon Titan Text G1 - Express.
    model_id = "amazon.titan-text-express-v1"

    # Define the message for the model.
    message = "Describe the purpose of a 'hello world' program in one line."

    # Send the message to the model.
    response = converse(brt, model_id, message)

    print(f"Response: {response}")

    logger.info("Done.")


if __name__ == "__main__":
    main()

• Einzelheiten zur API finden Sie InvokeModelin AWS SDK for Python (Boto3) API Reference.

Hallo Amazon Bedrock

Die folgenden Codebeispiele zeigen, wie Sie mit Amazon Bedrock beginnen können.

Go

SDK für Go V2

Anmerkung

Es gibt noch mehr dazu. GitHub Hier finden Sie das vollständige Beispiel und erfahren, wie Sie das AWS -Code-Beispiel- einrichten und ausführen.

package main

import (
	"context"
	"encoding/json"
	"flag"
	"fmt"
	"log"
	"os"
	"strings"

	"github.com/aws/aws-sdk-go-v2/aws"
	"github.com/aws/aws-sdk-go-v2/config"
	"github.com/aws/aws-sdk-go-v2/service/bedrockruntime"
)

// Each model provider defines their own individual request and response formats.
// For the format, ranges, and default values for the different models, refer to:
// https://docs.aws.amazon.com/bedrock/latest/userguide/model-parameters.html

type ClaudeRequest struct {
	Prompt            string `json:"prompt"`
	MaxTokensToSample int    `json:"max_tokens_to_sample"`
	// Omitting optional request parameters
}

type ClaudeResponse struct {
	Completion string `json:"completion"`
}

// main uses the AWS SDK for Go (v2) to create an Amazon Bedrock Runtime client
// and invokes Anthropic Claude 2 inside your account and the chosen region.
// This example uses the default settings specified in your shared credentials
// and config files.
func main() {

	region := flag.String("region", "us-east-1", "The AWS region")
	flag.Parse()

	fmt.Printf("Using AWS region: %s\n", *region)

	ctx := context.Background()
	sdkConfig, err := config.LoadDefaultConfig(ctx, config.WithRegion(*region))
	if err != nil {
		fmt.Println("Couldn't load default configuration. Have you set up your AWS account?")
		fmt.Println(err)
		return
	}

	client := bedrockruntime.NewFromConfig(sdkConfig)

	modelId := "anthropic.claude-v2"

	prompt := "Hello, how are you today?"

	// Anthropic Claude requires you to enclose the prompt as follows:
	prefix := "Human: "
	postfix := "\n\nAssistant:"
	wrappedPrompt := prefix + prompt + postfix

	request := ClaudeRequest{
		Prompt:            wrappedPrompt,
		MaxTokensToSample: 200,
	}

	body, err := json.Marshal(request)
	if err != nil {
		log.Panicln("Couldn't marshal the request: ", err)
	}

	result, err := client.InvokeModel(ctx, &bedrockruntime.InvokeModelInput{
		ModelId:     aws.String(modelId),
		ContentType: aws.String("application/json"),
		Body:        body,
	})

	if err != nil {
		errMsg := err.Error()
		if strings.Contains(errMsg, "no such host") {
			fmt.Printf("Error: The Bedrock service is not available in the selected region. Please double-check the service availability for your region at https://aws.amazon.com/about-aws/global-infrastructure/regional-product-services/.\n")
		} else if strings.Contains(errMsg, "Could not resolve the foundation model") {
			fmt.Printf("Error: Could not resolve the foundation model from model identifier: \"%v\". Please verify that the requested model exists and is accessible within the specified region.\n", modelId)
		} else {
			fmt.Printf("Error: Couldn't invoke Anthropic Claude. Here's why: %v\n", err)
		}
		os.Exit(1)
	}

	var response ClaudeResponse

	err = json.Unmarshal(result.Body, &response)

	if err != nil {
		log.Fatal("failed to unmarshal", err)
	}
	fmt.Println("Prompt:\n", prompt)
	fmt.Println("Response from Anthropic Claude:\n", response.Completion)
}

• Einzelheiten zur API finden Sie InvokeModelin der AWS SDK für Go API-Referenz.

JavaScript

SDK für JavaScript (v3)

Anmerkung

Es gibt noch mehr dazu GitHub. Hier finden Sie das vollständige Beispiel und erfahren, wie Sie das AWS -Code-Beispiel- einrichten und ausführen.

/**
 * @typedef {Object} Content
 * @property {string} text
 *
 * @typedef {Object} Usage
 * @property {number} input_tokens
 * @property {number} output_tokens
 *
 * @typedef {Object} ResponseBody
 * @property {Content[]} content
 * @property {Usage} usage
 */

import { fileURLToPath } from "node:url";
import {
  BedrockRuntimeClient,
  InvokeModelCommand,
} from "@aws-sdk/client-bedrock-runtime";

const AWS_REGION = "us-east-1";

const MODEL_ID = "anthropic.claude-3-haiku-20240307-v1:0";
const PROMPT = "Hi. In a short paragraph, explain what you can do.";

const hello = async () => {
  console.log("=".repeat(35));
  console.log("Welcome to the Amazon Bedrock demo!");
  console.log("=".repeat(35));

  console.log("Model: Anthropic Claude 3 Haiku");
  console.log(`Prompt: ${PROMPT}\n`);
  console.log("Invoking model...\n");

  // Create a new Bedrock Runtime client instance.
  const client = new BedrockRuntimeClient({ region: AWS_REGION });

  // Prepare the payload for the model.
  const payload = {
    anthropic_version: "bedrock-2023-05-31",
    max_tokens: 1000,
    messages: [{ role: "user", content: [{ type: "text", text: PROMPT }] }],
  };

  // Invoke Claude with the payload and wait for the response.
  const apiResponse = await client.send(
    new InvokeModelCommand({
      contentType: "application/json",
      body: JSON.stringify(payload),
      modelId: MODEL_ID,
    }),
  );

  // Decode and return the response(s)
  const decodedResponseBody = new TextDecoder().decode(apiResponse.body);
  /** @type {ResponseBody} */
  const responseBody = JSON.parse(decodedResponseBody);
  const responses = responseBody.content;

  if (responses.length === 1) {
    console.log(`Response: ${responses[0].text}`);
  } else {
    console.log("Haiku returned multiple responses:");
    console.log(responses);
  }

  console.log(`\nNumber of input tokens:   ${responseBody.usage.input_tokens}`);
  console.log(`Number of output tokens: ${responseBody.usage.output_tokens}`);
};

if (process.argv[1] === fileURLToPath(import.meta.url)) {
  await hello();
}

• Einzelheiten zur API finden Sie InvokeModelin der AWS SDK for JavaScript API-Referenz.

Python

SDK für Python (Boto3)

Anmerkung

Es gibt noch mehr dazu GitHub. Hier finden Sie das vollständige Beispiel und erfahren, wie Sie das AWS -Code-Beispiel- einrichten und ausführen.

Senden Sie mit der InvokeModel Operation eine Aufforderung an ein Modell.

"""
Uses the Amazon Bedrock runtime client InvokeModel operation to send a prompt to a model.
"""
import logging
import json
import boto3


from botocore.exceptions import ClientError


logging.basicConfig(level=logging.INFO)
logger = logging.getLogger(__name__)


def invoke_model(brt, model_id, prompt):
    """
    Invokes the specified model with the supplied prompt.
    param brt: A bedrock runtime boto3 client
    param model_id: The model ID for the model that you want to use.
    param prompt: The prompt that you want to send to the model.

    :return: The text response from the model.
    """

    # Format the request payload using the model's native structure.
    native_request = {
        "inputText": prompt,
        "textGenerationConfig": {
            "maxTokenCount": 512,
            "temperature": 0.5,
            "topP": 0.9
        }
    }

    # Convert the native request to JSON.
    request = json.dumps(native_request)

    try:
        # Invoke the model with the request.
        response = brt.invoke_model(modelId=model_id, body=request)

        # Decode the response body.
        model_response = json.loads(response["body"].read())

        # Extract and print the response text.
        response_text = model_response["results"][0]["outputText"]
        return response_text

    except (ClientError, Exception) as e:
        print(f"ERROR: Can't invoke '{model_id}'. Reason: {e}")
        raise


def main():
    """Entry point for the example. Uses the AWS SDK for Python (Boto3)
    to create an Amazon Bedrock runtime client. Then sends a prompt to a model
    in the region set in the callers profile and credentials.
    """

    # Create an Amazon Bedrock Runtime client.
    brt = boto3.client("bedrock-runtime")

    # Set the model ID, e.g., Amazon Titan Text G1 - Express.
    model_id = "amazon.titan-text-express-v1"

    # Define the prompt for the model.
    prompt = "Describe the purpose of a 'hello world' program in one line."

    # Send the prompt to the model.
    response = invoke_model(brt, model_id, prompt)

    print(f"Response: {response}")

    logger.info("Done.")


if __name__ == "__main__":
    main()

Senden Sie mit der Converse-Operation eine Benutzernachricht an ein Modell.

"""
Uses the Amazon Bedrock runtime client Converse operation to send a user message to a model.
"""
import logging
import boto3

from botocore.exceptions import ClientError


logging.basicConfig(level=logging.INFO)
logger = logging.getLogger(__name__)


def converse(brt, model_id, user_message):
    """
    Uses the Converse operation to send a user message to the supplied model.
    param brt: A bedrock runtime boto3 client
    param model_id: The model ID for the model that you want to use.
    param user message: The user message that you want to send to the model.

    :return: The text response from the model.
    """

    # Format the request payload using the model's native structure.
    conversation = [
    {
        "role": "user",
        "content": [{"text": user_message}],
    }
]

    try:
        # Send the message to the model, using a basic inference configuration.
        response = brt.converse(
            modelId=model_id,
            messages=conversation,
            inferenceConfig={"maxTokens": 512, "temperature": 0.5, "topP": 0.9},
        )

        # Extract and print the response text.
        response_text = response["output"]["message"]["content"][0]["text"]
        return response_text

    except (ClientError, Exception) as e:
        print(f"ERROR: Can't invoke '{model_id}'. Reason: {e}")
        raise


def main():
    """Entry point for the example. Uses the AWS SDK for Python (Boto3)
    to create an Amazon Bedrock runtime client. Then sends a user message to a model
    in the region set in the callers profile and credentials.
    """

    # Create an Amazon Bedrock Runtime client.
    brt = boto3.client("bedrock-runtime")

    # Set the model ID, e.g., Amazon Titan Text G1 - Express.
    model_id = "amazon.titan-text-express-v1"

    # Define the message for the model.
    message = "Describe the purpose of a 'hello world' program in one line."

    # Send the message to the model.
    response = converse(brt, model_id, message)

    print(f"Response: {response}")

    logger.info("Done.")


if __name__ == "__main__":
    main()

• Einzelheiten zur API finden Sie InvokeModelin AWS SDK for Python (Boto3) API Reference.

Codebeispiele

• Grundlagen

  • Hallo Amazon Bedrock

• Szenarien

  • Erstellen Sie eine Playground-Anwendung zur Interaktion mit Amazon Bedrock Foundation-Modellen

  • Rufen Sie mehrere Foundation-Modelle auf Amazon Bedrock auf

  • Orchestrieren Sie generative KI-Anwendungen mit Step Functions

  • Verwendung des Tools mit der Converse API

• AI21 Labs Jurassic-2

  • Converse

  • InvokeModel

• Amazon Nova

  • Converse

  • ConverseStream

  • Szenario: Verwendung des Tools mit der Converse API

• Amazon Nova Leinwand

  • InvokeModel

• Amazon Nova Reel

  • T ext-to-video

• Amazon Titan Image Generator

  • InvokeModel

• Amazon Titan Text

  • Converse

  • ConverseStream

  • InvokeModel

  • InvokeModelWithResponseStream

• Amazon Titan Text Embeddings

  • InvokeModel

• Anthropic Claude

  • Converse

  • ConverseStream

  • InvokeModel

  • InvokeModelWithResponseStream

  • Reasoning

  • Argumentation mit einer Streaming-Antwort

  • Szenario: Verwendung des Tools mit der Converse API

• Cohere Command

  • Converse

  • ConverseStream

  • InvokeModel: Befehl R und R+

  • InvokeModel: Command und Command Light

  • InvokeModelWithResponseStream: Befehl R und R+

  • InvokeModelWithResponseStream: Befehls- und Befehlsampel

  • Szenario: Verwendung des Tools mit der Converse API

• Meta Lama

  • Converse

  • ConverseStream

  • InvokeModel: Lama 3

  • InvokeModelWithResponseStream: Lama 3

• Mistral KI

  • Converse

  • ConverseStream

  • InvokeModel

  • InvokeModelWithResponseStream

• Stabile Diffusion

  • InvokeModel