Invoque Anthropic Claude en Amazon Bedrock mediante la API Invoke Model

Los siguientes ejemplos de código muestran cómo enviar un mensaje de texto a Anthropic Claude mediante la API Invoke Model.

.NET

AWS SDK for .NET

nota

Hay más información al respecto GitHub. Busque el ejemplo completo y aprenda a configurar y ejecutar en el Repositorio de ejemplos de código de AWS.

Usa la API Invoke Model para enviar un mensaje de texto.

// Use the native inference API to send a text message to Anthropic Claude.

using System;
using System.IO;
using System.Text.Json;
using System.Text.Json.Nodes;
using Amazon;
using Amazon.BedrockRuntime;
using Amazon.BedrockRuntime.Model;

// Create a Bedrock Runtime client in the AWS Region you want to use.
var client = new AmazonBedrockRuntimeClient(RegionEndpoint.USEast1);

// Set the model ID, e.g., Claude 3 Haiku.
var modelId = "anthropic.claude-3-haiku-20240307-v1:0";

// Define the user message.
var userMessage = "Describe the purpose of a 'hello world' program in one line.";

//Format the request payload using the model's native structure.
var nativeRequest = JsonSerializer.Serialize(new
{
    anthropic_version = "bedrock-2023-05-31",
    max_tokens = 512,
    temperature = 0.5,
    messages = new[]
    {
        new { role = "user", content = userMessage }
    }
});

// Create a request with the model ID and the model's native request payload.
var request = new InvokeModelRequest()
{
    ModelId = modelId,
    Body = new MemoryStream(System.Text.Encoding.UTF8.GetBytes(nativeRequest)),
    ContentType = "application/json"
};

try
{
    // Send the request to the Bedrock Runtime and wait for the response.
    var response = await client.InvokeModelAsync(request);

    // Decode the response body.
    var modelResponse = await JsonNode.ParseAsync(response.Body);

    // Extract and print the response text.
    var responseText = modelResponse["content"]?[0]?["text"] ?? "";
    Console.WriteLine(responseText);
}
catch (AmazonBedrockRuntimeException e)
{
    Console.WriteLine($"ERROR: Can't invoke '{modelId}'. Reason: {e.Message}");
    throw;
}

• Para obtener más información sobre la API, consulte InvokeModella referencia de AWS SDK for .NET la API.

Go

SDK para Go V2

nota

Hay más información al respecto GitHub. Busque el ejemplo completo y aprenda a configurar y ejecutar en el Repositorio de ejemplos de código de AWS.

Invoque el modelo fundacional Anthropic Claude 2 para generar texto.

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

type ClaudeRequest struct {
	Prompt            string   `json:"prompt"`
	MaxTokensToSample int      `json:"max_tokens_to_sample"`
	Temperature       float64  `json:"temperature,omitempty"`
	StopSequences     []string `json:"stop_sequences,omitempty"`
}

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

// Invokes Anthropic Claude on Amazon Bedrock to run an inference using the input
// provided in the request body.
func (wrapper InvokeModelWrapper) InvokeClaude(prompt string) (string, error) {
	modelId := "anthropic.claude-v2"

	// Anthropic Claude requires enclosing the prompt as follows:
	enclosedPrompt := "Human: " + prompt + "\n\nAssistant:"

	body, err := json.Marshal(ClaudeRequest{
		Prompt:            enclosedPrompt,
		MaxTokensToSample: 200,
		Temperature:       0.5,
		StopSequences:     []string{"\n\nHuman:"},
	})

	if err != nil {
		log.Fatal("failed to marshal", err)
	}

	output, err := wrapper.BedrockRuntimeClient.InvokeModel(context.TODO(), &bedrockruntime.InvokeModelInput{
		ModelId:     aws.String(modelId),
		ContentType: aws.String("application/json"),
		Body:        body,
	})

	if err != nil {
		ProcessError(err, modelId)
	}

	var response ClaudeResponse
	if err := json.Unmarshal(output.Body, &response); err != nil {
		log.Fatal("failed to unmarshal", err)
	}

	return response.Completion, nil
}

• Para obtener más información sobre la API, consulta InvokeModella Referencia AWS SDK for Go de la API.

Java

SDK para Java 2.x

nota

Hay más información al respecto GitHub. Busque el ejemplo completo y aprenda a configurar y ejecutar en el Repositorio de ejemplos de código de AWS.

Usa la API Invoke Model para enviar un mensaje de texto.

// Use the native inference API to send a text message to Anthropic Claude.

import org.json.JSONObject;
import org.json.JSONPointer;
import software.amazon.awssdk.auth.credentials.DefaultCredentialsProvider;
import software.amazon.awssdk.core.SdkBytes;
import software.amazon.awssdk.core.exception.SdkClientException;
import software.amazon.awssdk.regions.Region;
import software.amazon.awssdk.services.bedrockruntime.BedrockRuntimeClient;

public class InvokeModel {

    public static String invokeModel() {

        // Create a Bedrock Runtime client in the AWS Region you want to use.
        // Replace the DefaultCredentialsProvider with your preferred credentials provider.
        var client = BedrockRuntimeClient.builder()
                .credentialsProvider(DefaultCredentialsProvider.create())
                .region(Region.US_EAST_1)
                .build();

        // Set the model ID, e.g., Claude 3 Haiku.
        var modelId = "anthropic.claude-3-haiku-20240307-v1:0";

        // The InvokeModel API uses the model's native payload.
        // Learn more about the available inference parameters and response fields at:
        // https://docs.aws.amazon.com/bedrock/latest/userguide/model-parameters-anthropic-claude-messages.html
        var nativeRequestTemplate = """
                {
                    "anthropic_version": "bedrock-2023-05-31",
                    "max_tokens": 512,
                    "temperature": 0.5,
                    "messages": [{
                        "role": "user",
                        "content": "{{prompt}}"
                    }]
                }""";

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

        // Embed the prompt in the model's native request payload.
        String nativeRequest = nativeRequestTemplate.replace("{{prompt}}", prompt);

        try {
            // Encode and send the request to the Bedrock Runtime.
            var response = client.invokeModel(request -> request
                    .body(SdkBytes.fromUtf8String(nativeRequest))
                    .modelId(modelId)
            );

            // Decode the response body.
            var responseBody = new JSONObject(response.body().asUtf8String());

            // Retrieve the generated text from the model's response.
            var text = new JSONPointer("/content/0/text").queryFrom(responseBody).toString();
            System.out.println(text);

            return text;

        } catch (SdkClientException e) {
            System.err.printf("ERROR: Can't invoke '%s'. Reason: %s", modelId, e.getMessage());
            throw new RuntimeException(e);
        }
    }

    public static void main(String[] args) {
        invokeModel();
    }
}

• Para obtener más información sobre la API, consulte InvokeModella referencia de AWS SDK for Java 2.x la API.

JavaScript

SDK para JavaScript (v3)

nota

Hay más información. GitHub Busque el ejemplo completo y aprenda a configurar y ejecutar en el Repositorio de ejemplos de código de AWS.

Usa la API Invoke Model para enviar un mensaje de texto.

// Copyright Amazon.com, Inc. or its affiliates. All Rights Reserved.
// SPDX-License-Identifier: Apache-2.0

import { fileURLToPath } from "url";

import { FoundationModels } from "../../config/foundation_models.js";
import {
  BedrockRuntimeClient,
  InvokeModelCommand,
  InvokeModelWithResponseStreamCommand,
} from "@aws-sdk/client-bedrock-runtime";

/**
 * @typedef {Object} ResponseContent
 * @property {string} text
 *
 * @typedef {Object} MessagesResponseBody
 * @property {ResponseContent[]} content
 *
 * @typedef {Object} Delta
 * @property {string} text
 *
 * @typedef {Object} Message
 * @property {string} role
 *
 * @typedef {Object} Chunk
 * @property {string} type
 * @property {Delta} delta
 * @property {Message} message
 */

/**
 * Invokes Anthropic Claude 3 using the Messages API.
 *
 * To learn more about the Anthropic Messages API, go to:
 * https://docs.aws.amazon.com/bedrock/latest/userguide/model-parameters-anthropic-claude-messages.html
 *
 * @param {string} prompt - The input text prompt for the model to complete.
 * @param {string} [modelId] - The ID of the model to use. Defaults to "anthropic.claude-3-haiku-20240307-v1:0".
 */
export const invokeModel = async (
  prompt,
  modelId = "anthropic.claude-3-haiku-20240307-v1:0",
) => {
  // Create a new Bedrock Runtime client instance.
  const client = new BedrockRuntimeClient({ region: "us-east-1" });

  // 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 command = new InvokeModelCommand({
    contentType: "application/json",
    body: JSON.stringify(payload),
    modelId,
  });
  const apiResponse = await client.send(command);

  // Decode and return the response(s)
  const decodedResponseBody = new TextDecoder().decode(apiResponse.body);
  /** @type {MessagesResponseBody} */
  const responseBody = JSON.parse(decodedResponseBody);
  return responseBody.content[0].text;
};

/**
 * Invokes Anthropic Claude 3 and processes the response stream.
 *
 * To learn more about the Anthropic Messages API, go to:
 * https://docs.aws.amazon.com/bedrock/latest/userguide/model-parameters-anthropic-claude-messages.html
 *
 * @param {string} prompt - The input text prompt for the model to complete.
 * @param {string} [modelId] - The ID of the model to use. Defaults to "anthropic.claude-3-haiku-20240307-v1:0".
 */
export const invokeModelWithResponseStream = async (
  prompt,
  modelId = "anthropic.claude-3-haiku-20240307-v1:0",
) => {
  // Create a new Bedrock Runtime client instance.
  const client = new BedrockRuntimeClient({ region: "us-east-1" });

  // 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 API to respond.
  const command = new InvokeModelWithResponseStreamCommand({
    contentType: "application/json",
    body: JSON.stringify(payload),
    modelId,
  });
  const apiResponse = await client.send(command);

  let completeMessage = "";

  // Decode and process the response stream
  for await (const item of apiResponse.body) {
    /** @type Chunk */
    const chunk = JSON.parse(new TextDecoder().decode(item.chunk.bytes));
    const chunk_type = chunk.type;

    if (chunk_type === "content_block_delta") {
      const text = chunk.delta.text;
      completeMessage = completeMessage + text;
      process.stdout.write(text);
    }
  }

  // Return the final response
  return completeMessage;
};

// Invoke the function if this file was run directly.
if (process.argv[1] === fileURLToPath(import.meta.url)) {
  const prompt = 'Write a paragraph starting with: "Once upon a time..."';
  const modelId = FoundationModels.CLAUDE_3_HAIKU.modelId;
  console.log(`Prompt: ${prompt}`);
  console.log(`Model ID: ${modelId}`);

  try {
    console.log("-".repeat(53));
    const response = await invokeModel(prompt, modelId);
    console.log("\n" + "-".repeat(53));
    console.log("Final structured response:");
    console.log(response);
  } catch (err) {
    console.log(`\n${err}`);
  }
}

• Para obtener más información sobre la API, consulte InvokeModella referencia de AWS SDK for JavaScript la API.

PHP

SDK para PHP

nota

Hay más información al respecto GitHub. Busque el ejemplo completo y aprenda a configurar y ejecutar en el Repositorio de ejemplos de código de AWS.

Invoque el modelo fundacional Anthropic Claude 2 para generar texto.

    public function invokeClaude($prompt)
    {
        # The different model providers have individual request and response formats.
        # For the format, ranges, and default values for Anthropic Claude, refer to:
        # https://docs.aws.amazon.com/bedrock/latest/userguide/model-parameters-claude.html

        $completion = "";

        try {
            $modelId = 'anthropic.claude-v2';

            # Claude requires you to enclose the prompt as follows:
            $prompt = "\n\nHuman: {$prompt}\n\nAssistant:";

            $body = [
                'prompt' => $prompt,
                'max_tokens_to_sample' => 200,
                'temperature' => 0.5,
                'stop_sequences' => ["\n\nHuman:"],
            ];

            $result = $this->bedrockRuntimeClient->invokeModel([
                'contentType' => 'application/json',
                'body' => json_encode($body),
                'modelId' => $modelId,
            ]);

            $response_body = json_decode($result['body']);

            $completion = $response_body->completion;
        } catch (Exception $e) {
            echo "Error: ({$e->getCode()}) - {$e->getMessage()}\n";
        }

        return $completion;
    }

• Para obtener más información sobre la API, consulta InvokeModella Referencia AWS SDK for PHP de la API.

Python

SDK para Python (Boto3)

nota

Hay más información al respecto GitHub. Busque el ejemplo completo y aprenda a configurar y ejecutar en el Repositorio de ejemplos de código de AWS.

Usa la API Invoke Model para enviar un mensaje de texto.

# Use the native inference API to send a text message to Anthropic Claude.

import boto3
import json

from botocore.exceptions import ClientError

# Create a Bedrock Runtime client in the AWS Region of your choice.
client = boto3.client("bedrock-runtime", region_name="us-east-1")

# Set the model ID, e.g., Claude 3 Haiku.
model_id = "anthropic.claude-3-haiku-20240307-v1:0"

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

# Format the request payload using the model's native structure.
native_request = {
    "anthropic_version": "bedrock-2023-05-31",
    "max_tokens": 512,
    "temperature": 0.5,
    "messages": [
        {
            "role": "user",
            "content": [{"type": "text", "text": prompt}],
        }
    ],
}

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

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

except (ClientError, Exception) as e:
    print(f"ERROR: Can't invoke '{model_id}'. Reason: {e}")
    exit(1)

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

# Extract and print the response text.
response_text = model_response["content"][0]["text"]
print(response_text)

• Para obtener más información sobre la API, consulta InvokeModella AWS Referencia de API de SDK for Python (Boto3).

SAP ABAP

SDK de SAP ABAP

nota

Hay más información al respecto. GitHub Busque el ejemplo completo y aprenda a configurar y ejecutar en el Repositorio de ejemplos de código de AWS.

Invoque el modelo fundacional Anthropic Claude 2 para generar texto. En este ejemplo, se utilizan funciones de /US2/CL_JSON que podrían no estar disponibles en algunas versiones. NetWeaver

    "Claude V2 Input Parameters should be in a format like this:
*   {
*     "prompt":"\n\nHuman:\\nTell me a joke\n\nAssistant:\n",
*     "max_tokens_to_sample":2048,
*     "temperature":0.5,
*     "top_k":250,
*     "top_p":1.0,
*     "stop_sequences":[]
*   }

    DATA: BEGIN OF ls_input,
            prompt               TYPE string,
            max_tokens_to_sample TYPE /aws1/rt_shape_integer,
            temperature          TYPE /aws1/rt_shape_float,
            top_k                TYPE /aws1/rt_shape_integer,
            top_p                TYPE /aws1/rt_shape_float,
            stop_sequences       TYPE /aws1/rt_stringtab,
          END OF ls_input.

    "Leave ls_input-stop_sequences empty.
    ls_input-prompt = |\n\nHuman:\\n{ iv_prompt }\n\nAssistant:\n|.
    ls_input-max_tokens_to_sample = 2048.
    ls_input-temperature = '0.5'.
    ls_input-top_k = 250.
    ls_input-top_p = 1.

    "Serialize into JSON with /ui2/cl_json -- this assumes SAP_UI is installed.
    DATA(lv_json) = /ui2/cl_json=>serialize(
      data = ls_input
                pretty_name   = /ui2/cl_json=>pretty_mode-low_case ).

    TRY.
        DATA(lo_response) = lo_bdr->invokemodel(
          iv_body = /aws1/cl_rt_util=>string_to_xstring( lv_json )
          iv_modelid = 'anthropic.claude-v2'
          iv_accept = 'application/json'
          iv_contenttype = 'application/json' ).

        "Claude V2 Response format will be:
*       {
*         "completion": "Knock Knock...",
*         "stop_reason": "stop_sequence"
*       }
        DATA: BEGIN OF ls_response,
                completion  TYPE string,
                stop_reason TYPE string,
              END OF ls_response.

        /ui2/cl_json=>deserialize(
          EXPORTING jsonx = lo_response->get_body( )
                    pretty_name = /ui2/cl_json=>pretty_mode-camel_case
          CHANGING  data  = ls_response ).

        DATA(lv_answer) = ls_response-completion.
      CATCH /aws1/cx_bdraccessdeniedex INTO DATA(lo_ex).
        WRITE / lo_ex->get_text( ).
        WRITE / |Don't forget to enable model access at https://console.aws.amazon.com/bedrock/home?#/modelaccess|.

    ENDTRY.

Invoque el modelo básico de Anthropic Claude 2 para generar texto utilizando el cliente de alto nivel L2.

    TRY.
        DATA(lo_bdr_l2_claude) = /aws1/cl_bdr_l2_factory=>create_claude_2( lo_bdr ).
        " iv_prompt can contain a prompt like 'tell me a joke about Java programmers'.
        DATA(lv_answer) = lo_bdr_l2_claude->prompt_for_text( iv_prompt ).
      CATCH /aws1/cx_bdraccessdeniedex INTO DATA(lo_ex).
        WRITE / lo_ex->get_text( ).
        WRITE / |Don't forget to enable model access at https://console.aws.amazon.com/bedrock/home?#/modelaccess|.

    ENDTRY.

• Para obtener más información sobre la API, consulte el AWS SDK para ver InvokeModella referencia sobre la API ABAP de SAP.

Para obtener una lista completa de guías para desarrolladores del AWS SDK y ejemplos de código, consulteUso de este servicio con un AWS SDK. En este tema también se incluye información sobre cómo comenzar a utilizar el SDK y detalles sobre sus versiones anteriores.