Le traduzioni sono generate tramite traduzione automatica. In caso di conflitto tra il contenuto di una traduzione e la versione originale in Inglese, quest'ultima prevarrà.
# Esempi di codice di Amazon Nova
**Nota**
Questa documentazione è per Amazon Nova versione 1. Per esempi di codice relativi ad Amazon Nova 2, visita [Code library](https://docs.aws.amazon.com/nova/latest/nova2-userguide/code-library.html).
Gli esempi di codice seguenti illustrano alcuni casi d’uso comuni per Amazon Nova.
**Topics**
+ [Inviare un messaggio con l’API Converse](code-examples-converse.md)
+ [Invia un messaggio con l' ConverseStream API](code-examples-conversestream.md)
+ [Generare un’immagine](code-examples-image.md)
+ [Generare un video](code-examples-video.md)
+ [Utilizzare uno strumento con Amazon Nova](code-examples-tool.md)
# Inviare un messaggio con l’API Converse
Gli esempi di codice seguenti mostrano come inviare un messaggio di testo ad Amazon Nova utilizzando l’API Converse di Bedrock.
------
#### [ .NET ]
**SDK per .NET**
C'è altro su GitHub. Trova l'esempio completo e scopri di più sulla configurazione e l'esecuzione nel [Repository di esempi di codice AWS](https://github.com/awsdocs/aws-doc-sdk-examples/tree/main/dotnetv3/Bedrock-runtime#code-examples).
Invia un messaggio di testo ad Amazon Nova utilizzando l’API Converse di Bedrock.
```
// Use the Converse API to send a text message to Amazon Nova.
using System;
using System.Collections.Generic;
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., Amazon Nova Lite.
var modelId = "amazon.nova-lite-v1:0";
// Define the user message.
var userMessage = "Describe the purpose of a 'hello world' program in one line.";
// Create a request with the model ID, the user message, and an inference configuration.
var request = new ConverseRequest
{
ModelId = modelId,
Messages = new List
{
new Message
{
Role = ConversationRole.User,
Content = new List { new ContentBlock { Text = userMessage } }
}
},
InferenceConfig = new InferenceConfiguration()
{
MaxTokens = 512,
Temperature = 0.5F,
TopP = 0.9F
}
};
try
{
// Send the request to the Bedrock Runtime and wait for the result.
var response = await client.ConverseAsync(request);
// Extract and print the response text.
string responseText = response?.Output?.Message?.Content?[0]?.Text ?? "";
Console.WriteLine(responseText);
}
catch (AmazonBedrockRuntimeException e)
{
Console.WriteLine($"ERROR: Can't invoke '{modelId}'. Reason: {e.Message}");
throw;
}
```
Invia una conversazione di messaggi ad Amazon Nova utilizzando l’API Converse di Bedrock con una configurazione dello strumento.
```
///
/// Wrapper class for interacting with the Amazon Bedrock Converse API.
///
public class BedrockActionsWrapper
{
private readonly IAmazonBedrockRuntime _bedrockClient;
private readonly ILogger _logger;
///
/// Initializes a new instance of the class.
///
/// The Bedrock Converse API client.
/// The logger instance.
public BedrockActionsWrapper(IAmazonBedrockRuntime bedrockClient, ILogger logger)
{
_bedrockClient = bedrockClient;
_logger = logger;
}
///
/// Sends a Converse request to the Amazon Bedrock Converse API.
///
/// The Bedrock Model Id.
/// A system prompt instruction.
/// The array of messages in the conversation.
/// The specification for a tool.
/// The response of the model.
public async Task SendConverseRequestAsync(string modelId, string systemPrompt, List conversation, ToolSpecification toolSpec)
{
try
{
var request = new ConverseRequest()
{
ModelId = modelId,
System = new List()
{
new SystemContentBlock()
{
Text = systemPrompt
}
},
Messages = conversation,
ToolConfig = new ToolConfiguration()
{
Tools = new List()
{
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;
}
}
}
```
+ Per informazioni dettagliate sull’API, consulta [Converse](https://docs.aws.amazon.com/goto/DotNetSDKV3/bedrock-runtime-2023-09-30/Converse) nella *documentazione di riferimento dell’API AWS SDK per .NET *.
------
#### [ Java ]
**SDK per Java 2.x**
C'è dell'altro GitHub. Trova l'esempio completo e scopri di più sulla configurazione e l'esecuzione nel [Repository di esempi di codice AWS](https://github.com/awsdocs/aws-doc-sdk-examples/tree/main/javav2/example_code/bedrock-runtime#code-examples).
Invia un messaggio di testo ad Amazon Nova utilizzando l’API Converse di Bedrock con il client Java asincrono.
```
import software.amazon.awssdk.auth.credentials.DefaultCredentialsProvider;
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;
/**
* This example demonstrates how to use the Amazon Nova foundation models
* with an asynchronous Amazon Bedrock runtime client to generate text.
* It shows how to:
* - Set up the Amazon Bedrock runtime client
* - Create a message
* - Configure and send a request
* - Process the response
*/
public class ConverseAsync {
public static String converseAsync() {
// Step 1: Create the Amazon Bedrock runtime client
// The runtime client handles the communication with AI models on Amazon Bedrock
BedrockRuntimeAsyncClient client = BedrockRuntimeAsyncClient.builder()
.credentialsProvider(DefaultCredentialsProvider.create())
.region(Region.US_EAST_1)
.build();
// Step 2: Specify which model to use
// Available Amazon Nova models and their characteristics:
// - Amazon Nova Micro: Text-only model optimized for lowest latency and cost
// - Amazon Nova Lite: Fast, low-cost multimodal model for image, video, and text
// - Amazon Nova Pro: Advanced multimodal model balancing accuracy, speed, and cost
//
// For the latest available models, see:
// https://docs.aws.amazon.com/bedrock/latest/userguide/models-supported.html
String modelId = "amazon.nova-lite-v1:0";
// Step 3: Create the message
// The message includes the text prompt and specifies that it comes from the user
var inputText = "Describe the purpose of a 'hello world' program in one line.";
var message = Message.builder()
.content(ContentBlock.fromText(inputText))
.role(ConversationRole.USER)
.build();
// Step 4: Configure the request
// Optional parameters to control the model's response:
// - maxTokens: maximum number of tokens to generate
// - temperature: randomness (max: 1.0, default: 0.7)
// OR
// - topP: diversity of word choice (max: 1.0, default: 0.9)
// Note: Use either temperature OR topP, but not both
ConverseRequest request = ConverseRequest.builder()
.modelId(modelId)
.messages(message)
.inferenceConfig(config -> config
.maxTokens(500) // The maximum response length
.temperature(0.5F) // Using temperature for randomness control
//.topP(0.9F) // Alternative: use topP instead of temperature
).build();
// Step 5: Send and process the request asynchronously
// - Send the request to the model
// - Extract and return the generated text from the response
try {
CompletableFuture asyncResponse = client.converse(request);
return asyncResponse.thenApply(
response -> response.output().message().content().get(0).text()
).get();
} catch (Exception e) {
System.err.printf("Can't invoke '%s': %s", modelId, e.getMessage());
throw new RuntimeException(e);
}
}
public static void main(String[] args) {
String response = converseAsync();
System.out.println(response);
}
}
```
Invia un messaggio di testo ad Amazon Nova utilizzando l’API Converse di Bedrock.
```
import software.amazon.awssdk.auth.credentials.DefaultCredentialsProvider;
import software.amazon.awssdk.core.exception.SdkClientException;
import software.amazon.awssdk.regions.Region;
import software.amazon.awssdk.services.bedrockruntime.BedrockRuntimeClient;
import software.amazon.awssdk.services.bedrockruntime.model.*;
/**
* This example demonstrates how to use the Amazon Nova foundation models
* with a synchronous Amazon Bedrock runtime client to generate text.
* It shows how to:
* - Set up the Amazon Bedrock runtime client
* - Create a message
* - Configure and send a request
* - Process the response
*/
public class Converse {
public static String converse() {
// Step 1: Create the Amazon Bedrock runtime client
// The runtime client handles the communication with AI models on Amazon Bedrock
BedrockRuntimeClient client = BedrockRuntimeClient.builder()
.credentialsProvider(DefaultCredentialsProvider.create())
.region(Region.US_EAST_1)
.build();
// Step 2: Specify which model to use
// Available Amazon Nova models and their characteristics:
// - Amazon Nova Micro: Text-only model optimized for lowest latency and cost
// - Amazon Nova Lite: Fast, low-cost multimodal model for image, video, and text
// - Amazon Nova Pro: Advanced multimodal model balancing accuracy, speed, and cost
//
// For the latest available models, see:
// https://docs.aws.amazon.com/bedrock/latest/userguide/models-supported.html
String modelId = "amazon.nova-lite-v1:0";
// Step 3: Create the message
// The message includes the text prompt and specifies that it comes from the user
var inputText = "Describe the purpose of a 'hello world' program in one line.";
var message = Message.builder()
.content(ContentBlock.fromText(inputText))
.role(ConversationRole.USER)
.build();
// Step 4: Configure the request
// Optional parameters to control the model's response:
// - maxTokens: maximum number of tokens to generate
// - temperature: randomness (max: 1.0, default: 0.7)
// OR
// - topP: diversity of word choice (max: 1.0, default: 0.9)
// Note: Use either temperature OR topP, but not both
ConverseRequest request = ConverseRequest.builder()
.modelId(modelId)
.messages(message)
.inferenceConfig(config -> config
.maxTokens(500) // The maximum response length
.temperature(0.5F) // Using temperature for randomness control
//.topP(0.9F) // Alternative: use topP instead of temperature
).build();
// Step 5: Send and process the request
// - Send the request to the model
// - Extract and return the generated text from the response
try {
ConverseResponse response = client.converse(request);
return response.output().message().content().get(0).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) {
String response = converse();
System.out.println(response);
}
}
```
+ Per informazioni dettagliate sull’API, consulta [Converse](https://docs.aws.amazon.com/goto/SdkForJavaV2/bedrock-runtime-2023-09-30/Converse) nella *documentazione di riferimento dell’API AWS SDK for Java 2.x *.
------
#### [ JavaScript ]
**SDK per JavaScript (v3)**
C'è altro da fare. GitHub Trova l'esempio completo e scopri di più sulla configurazione e l'esecuzione nel [Repository di esempi di codice AWS](https://github.com/awsdocs/aws-doc-sdk-examples/tree/main/javascriptv3/example_code/bedrock-runtime#code-examples).
Invia un messaggio di testo ad Amazon Nova utilizzando l’API Converse di Bedrock.
```
// This example demonstrates how to use the Amazon Nova foundation models to generate text.
// It shows how to:
// - Set up the Amazon Bedrock runtime client
// - Create a message
// - Configure and send a request
// - Process the response
import {
BedrockRuntimeClient,
ConversationRole,
ConverseCommand,
} from "@aws-sdk/client-bedrock-runtime";
// Step 1: Create the Amazon Bedrock runtime client
// Credentials will be automatically loaded from the environment.
const client = new BedrockRuntimeClient({ region: "us-east-1" });
// Step 2: Specify which model to use:
// Available Amazon Nova models and their characteristics:
// - Amazon Nova Micro: Text-only model optimized for lowest latency and cost
// - Amazon Nova Lite: Fast, low-cost multimodal model for image, video, and text
// - Amazon Nova Pro: Advanced multimodal model balancing accuracy, speed, and cost
//
// For the most current model IDs, see:
// https://docs.aws.amazon.com/bedrock/latest/userguide/models-supported.html
const modelId = "amazon.nova-lite-v1:0";
// Step 3: Create the message
// The message includes the text prompt and specifies that it comes from the user
const inputText =
"Describe the purpose of a 'hello world' program in one line.";
const message = {
content: [{ text: inputText }],
role: ConversationRole.USER,
};
// Step 4: Configure the request
// Optional parameters to control the model's response:
// - maxTokens: maximum number of tokens to generate
// - temperature: randomness (max: 1.0, default: 0.7)
// OR
// - topP: diversity of word choice (max: 1.0, default: 0.9)
// Note: Use either temperature OR topP, but not both
const request = {
modelId,
messages: [message],
inferenceConfig: {
maxTokens: 500, // The maximum response length
temperature: 0.5, // Using temperature for randomness control
//topP: 0.9, // Alternative: use topP instead of temperature
},
};
// Step 5: Send and process the request
// - Send the request to the model
// - Extract and return the generated text from the response
try {
const response = await client.send(new ConverseCommand(request));
console.log(response.output.message.content[0].text);
} catch (error) {
console.error(`ERROR: Can't invoke '${modelId}'. Reason: ${error.message}`);
throw error;
}
```
Invia una conversazione di messaggi ad Amazon Nova utilizzando l’API Converse di Bedrock con una configurazione dello strumento.
```
// This example demonstrates how to send a conversation of messages to Amazon Nova using Bedrock's Converse API with a tool configuration.
// It shows how to:
// - 1. Set up the Amazon Bedrock runtime client
// - 2. Define the parameters required enable Amazon Bedrock to use a tool when formulating its response (model ID, user input, system prompt, and the tool spec)
// - 3. Send the request to Amazon Bedrock, and returns the response.
// - 4. Add the tool response to the conversation, and send it back to Amazon Bedrock.
// - 5. Publish the response.
import {
BedrockRuntimeClient,
ConverseCommand,
} from "@aws-sdk/client-bedrock-runtime";
// Step 1: Create the Amazon Bedrock runtime client
// Credentials will be automatically loaded from the environment
const bedRockRuntimeClient = new BedrockRuntimeClient({
region: "us-east-1",
});
// Step 2. Define the parameters required enable Amazon Bedrock to use a tool when formulating its response.
// The Bedrock Model ID.
const modelId = "amazon.nova-lite-v1:0";
// The system prompt to help Amazon Bedrock craft it's response.
const system_prompt = [
{
text:
"You are a music expert that provides the most popular song played on a radio station, using only the\n" +
"the top_song tool, which he call sign for the radio station for which you want the most popular song. " +
"Example calls signs are WZPZ and WKRP. \n" +
"- Only use the top_song tool. Never guess or make up information. \n" +
"- If the tool errors, apologize, explain weather is unavailable, and suggest other options.\n" +
"- Only respond to queries about the most popular song played on a radio station\n" +
"Remind off-topic users of your purpose. \n" +
"- Never claim to search online, access external data, or use tools besides the top_song tool.\n",
},
];
// The user's question.
const message = [
{
role: "user",
content: [{ text: "What is the most popular song on WZPZ?" }],
},
];
// The tool specification. In this case, it uses an example schema for
// a tool that gets the most popular song played on a radio station.
const tool_config = {
tools: [
{
toolSpec: {
name: "top_song",
description: "Get the most popular song played on a radio station.",
inputSchema: {
json: {
type: "object",
properties: {
sign: {
type: "string",
description:
"The call sign for the radio station for which you want the most popular song. Example calls signs are WZPZ and WKRP.",
},
},
required: ["sign"],
},
},
},
},
],
};
// Helper function to return the song and artist from top_song tool.
async function get_top_song(call_sign) {
try {
if (call_sign === "WZPZ") {
const song = "Elemental Hotel";
const artist = "8 Storey Hike";
return { song, artist };
}
} catch (error) {
console.log(`${error.message}`);
}
}
// 3. Send the request to Amazon Bedrock, and returns the response.
export async function SendConversationtoBedrock(
modelId,
message,
system_prompt,
tool_config,
) {
try {
const response = await bedRockRuntimeClient.send(
new ConverseCommand({
modelId: modelId,
messages: message,
system: system_prompt,
toolConfig: tool_config,
}),
);
if (response.stopReason === "tool_use") {
const toolResultFinal = [];
try {
const output_message = response.output.message;
message.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 sign = toolUse.input.sign;
const toolUseID = toolUse.toolUseId;
console.log(
`Requesting tool ${toolUse.name}, Tool use id ${toolUseID}`,
);
if (toolUse.name === "top_song") {
const toolResult = [];
try {
const top_song = await get_top_song(toolUse.input.sign).then(
(top_song) => top_song,
);
const toolResult = {
toolResult: {
toolUseId: toolUseID,
content: [
{
json: { song: top_song.song, artist: top_song.artist },
},
],
},
};
toolResultFinal.push(toolResult);
} catch (err) {
const toolResult = {
toolUseId: toolUseID,
content: [{ json: { text: err.message } }],
status: "error",
};
}
}
}
}
const toolResultMessage = {
role: "user",
content: toolResultFinal,
};
// Step 4. Add the tool response to the conversation, and send it back to Amazon Bedrock.
message.push(toolResultMessage);
await SendConversationtoBedrock(
modelId,
message,
system_prompt,
tool_config,
);
} catch (caught) {
console.error(`${caught.message}`);
throw caught;
}
}
// 4. Publish the response.
if (response.stopReason === "end_turn") {
const finalMessage = response.output.message.content[0].text;
const messageToPrint = finalMessage.replace(/<[^>]+>/g);
console.log(messageToPrint.replace(/<[^>]+>/g));
return messageToPrint;
}
} 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;
}
}
}
await SendConversationtoBedrock(modelId, message, system_prompt, tool_config);
```
+ Per informazioni dettagliate sull’API, consulta [Converse](https://docs.aws.amazon.com/AWSJavaScriptSDK/v3/latest/client/bedrock-runtime/command/ConverseCommand) nella *documentazione di riferimento dell’API AWS SDK per JavaScript *.
------
#### [ Kotlin ]
**SDK per Kotlin**
C'è dell'altro GitHub. Trova l'esempio completo e scopri di più sulla configurazione e l'esecuzione nel [Repository di esempi di codice AWS](https://github.com/awsdocs/aws-doc-sdk-examples/tree/main/kotlin/services/bedrock-runtime#code-examples).
Invia un messaggio di testo ad Amazon Nova utilizzando l’API Converse di Bedrock.
```
import aws.sdk.kotlin.services.bedrockruntime.BedrockRuntimeClient
import aws.sdk.kotlin.services.bedrockruntime.model.ContentBlock
import aws.sdk.kotlin.services.bedrockruntime.model.ConversationRole
import aws.sdk.kotlin.services.bedrockruntime.model.ConverseRequest
import aws.sdk.kotlin.services.bedrockruntime.model.Message
/**
* This example demonstrates how to use the Amazon Nova foundation models to generate text.
* It shows how to:
* - Set up the Amazon Bedrock runtime client
* - Create a message
* - Configure and send a request
* - Process the response
*/
suspend fun main() {
converse().also { println(it) }
}
suspend fun converse(): String {
// Create and configure the Bedrock runtime client
BedrockRuntimeClient { region = "us-east-1" }.use { client ->
// Specify the model ID. For the latest available models, see:
// https://docs.aws.amazon.com/bedrock/latest/userguide/models-supported.html
val modelId = "amazon.nova-lite-v1:0"
// Create the message with the user's prompt
val prompt = "Describe the purpose of a 'hello world' program in one line."
val message = Message {
role = ConversationRole.User
content = listOf(ContentBlock.Text(prompt))
}
// Configure the request with optional model parameters
val request = ConverseRequest {
this.modelId = modelId
messages = listOf(message)
inferenceConfig {
maxTokens = 500 // Maximum response length
temperature = 0.5F // Lower values: more focused output
// topP = 0.8F // Alternative to temperature
}
}
// Send the request and process the model's response
runCatching {
val response = client.converse(request)
return response.output!!.asMessage().content.first().asText()
}.getOrElse { error ->
error.message?.let { e -> System.err.println("ERROR: Can't invoke '$modelId'. Reason: $e") }
throw RuntimeException("Failed to generate text with model $modelId", error)
}
}
}
```
+ Per informazioni dettagliate sull’API, consulta [Converse](https://sdk.amazonaws.com/kotlin/api/latest/index.html) nella *documentazione di riferimento dell’API AWS SDK per Kotlin*.
------
#### [ PHP ]
**SDK per PHP**
C'è dell'altro GitHub. Trova l'esempio completo e scopri di più sulla configurazione e l'esecuzione nel [Repository di esempi di codice AWS](https://github.com/awsdocs/aws-doc-sdk-examples/tree/main/php/example_code/bedrock-runtime#code-examples).
Invia un messaggio di testo ad Amazon Nova utilizzando l’API Converse di Bedrock.
```
// Use the Conversation API to send a text message to Amazon Nova.
use Aws\BedrockRuntime\BedrockRuntimeClient;
use Aws\Exception\AwsException;
use RuntimeException;
class Converse
{
public function converse(): string
{
// Create a Bedrock Runtime client in the AWS Region you want to use.
$client = new BedrockRuntimeClient([
'region' => 'us-east-1',
'profile' => 'default'
]);
// Set the model ID, e.g., Amazon Nova Lite.
$modelId = 'amazon.nova-lite-v1:0';
// Start a conversation with the user message.
$userMessage = "Describe the purpose of a 'hello world' program in one line.";
$conversation = [
[
"role" => "user",
"content" => [["text" => $userMessage]]
]
];
try {
// Send the message to the model, using a basic inference configuration.
$response = $client->converse([
'modelId' => $modelId,
'messages' => $conversation,
'inferenceConfig' => [
'maxTokens' => 512,
'temperature' => 0.5
]
]);
// Extract and return the response text.
$responseText = $response['output']['message']['content'][0]['text'];
return $responseText;
} catch (AwsException $e) {
echo "ERROR: Can't invoke {$modelId}. Reason: {$e->getAwsErrorMessage()}";
throw new RuntimeException("Failed to invoke model: " . $e->getAwsErrorMessage(), 0, $e);
}
}
}
$demo = new Converse();
echo $demo->converse();
```
+ Per informazioni dettagliate sull’API, consulta [Converse](https://docs.aws.amazon.com/goto/SdkForPHPV3/bedrock-runtime-2023-09-30/Converse) nella *documentazione di riferimento dell’API AWS SDK per PHP *.
------
#### [ Python ]
**SDK per Python (Boto3)**
C'è dell'altro GitHub. Trova l'esempio completo e scopri di più sulla configurazione e l'esecuzione nel [Repository di esempi di codice AWS](https://github.com/awsdocs/aws-doc-sdk-examples/tree/main/python/example_code/bedrock-runtime#code-examples).
Invia un messaggio di testo ad Amazon Nova utilizzando l’API Converse di Bedrock.
```
# Use the Conversation API to send a text message to Amazon Nova.
import boto3
from botocore.exceptions import ClientError
# Create a Bedrock Runtime client in the AWS Region you want to use.
client = boto3.client("bedrock-runtime", region_name="us-east-1")
# Set the model ID, e.g., Amazon Nova Lite.
model_id = "amazon.nova-lite-v1:0"
# Start a conversation with the user message.
user_message = "Describe the purpose of a 'hello world' program in one line."
conversation = [
{
"role": "user",
"content": [{"text": user_message}],
}
]
try:
# Send the message to the model, using a basic inference configuration.
response = client.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"]
print(response_text)
except (ClientError, Exception) as e:
print(f"ERROR: Can't invoke '{model_id}'. Reason: {e}")
exit(1)
```
+ Per informazioni dettagliate sull’API, consulta [Converse](https://docs.aws.amazon.com/goto/boto3/bedrock-runtime-2023-09-30/Converse) nella *documentazione di riferimento dell’API AWS SDK per Python (Boto3)*.
------
#### [ Swift ]
**SDK per Swift**
C'è altro da fare. GitHub Trova l'esempio completo e scopri di più sulla configurazione e l'esecuzione nel [Repository di esempi di codice AWS](https://github.com/awsdocs/aws-doc-sdk-examples/tree/main/swift/example_code/bedrock-runtime#code-examples).
Invia un messaggio di testo ad Amazon Nova utilizzando l’API Converse di Bedrock.
```
// An example demonstrating how to use the Conversation API to send
// a text message to Amazon Nova.
import AWSBedrockRuntime
func converse(_ textPrompt: String) async throws -> String {
// Create a Bedrock Runtime client in the AWS Region you want to use.
let config =
try await BedrockRuntimeClient.BedrockRuntimeClientConfiguration(
region: "us-east-1"
)
let client = BedrockRuntimeClient(config: config)
// Set the model ID.
let modelId = "amazon.nova-micro-v1:0"
// Start a conversation with the user message.
let message = BedrockRuntimeClientTypes.Message(
content: [.text(textPrompt)],
role: .user
)
// Optionally use inference parameters
let inferenceConfig =
BedrockRuntimeClientTypes.InferenceConfiguration(
maxTokens: 512,
stopSequences: ["END"],
temperature: 0.5,
topp: 0.9
)
// Create the ConverseInput to send to the model
let input = ConverseInput(
inferenceConfig: inferenceConfig, messages: [message], modelId: modelId)
// Send the ConverseInput to the model
let response = try await client.converse(input: input)
// Extract and return the response text.
if case let .message(msg) = response.output {
if case let .text(textResponse) = msg.content![0] {
return textResponse
} else {
return "No text response found in message content"
}
} else {
return "No message found in converse output"
}
}
```
+ Per informazioni dettagliate sull’API, consulta [Converse](https://sdk.amazonaws.com/swift/api/awsbedrockruntime/latest/documentation/awsbedrockruntime/bedrockruntimeclient/converse(input:)) nella *documentazione di riferimento dell’API AWS SDK per Swift*.
------
# Invia un messaggio con l' ConverseStream API
Gli esempi di codice seguenti mostrano come inviare un messaggio di testo ad Amazon Nova utilizzando l’API Converse di Bedrock ed elaborare il flusso di risposta in tempo reale.
------
#### [ .NET ]
**SDK per .NET**
C'è altro da fare GitHub. Trova l'esempio completo e scopri di più sulla configurazione e l'esecuzione nel [Repository di esempi di codice AWS](https://github.com/awsdocs/aws-doc-sdk-examples/tree/main/dotnetv3/Bedrock-runtime#code-examples).
Invia un messaggio di testo ad Amazon Nova utilizzando l’API Converse di Bedrock ed elabora il flusso di risposta in tempo reale.
```
// Use the Converse API to send a text message to Amazon Nova
// and print the response stream.
using System;
using System.Collections.Generic;
using System.Linq;
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., Amazon Nova Lite.
var modelId = "amazon.nova-lite-v1:0";
// Define the user message.
var userMessage = "Describe the purpose of a 'hello world' program in one line.";
// Create a request with the model ID, the user message, and an inference configuration.
var request = new ConverseStreamRequest
{
ModelId = modelId,
Messages = new List
{
new Message
{
Role = ConversationRole.User,
Content = new List { new ContentBlock { Text = userMessage } }
}
},
InferenceConfig = new InferenceConfiguration()
{
MaxTokens = 512,
Temperature = 0.5F,
TopP = 0.9F
}
};
try
{
// Send the request to the Bedrock Runtime and wait for the result.
var response = await client.ConverseStreamAsync(request);
// Extract and print the streamed response text in real-time.
foreach (var chunk in response.Stream.AsEnumerable())
{
if (chunk is ContentBlockDeltaEvent)
{
Console.Write((chunk as ContentBlockDeltaEvent).Delta.Text);
}
}
}
catch (AmazonBedrockRuntimeException e)
{
Console.WriteLine($"ERROR: Can't invoke '{modelId}'. Reason: {e.Message}");
throw;
}
```
+ Per i dettagli sull'API, consulta la [ConverseStream](https://docs.aws.amazon.com/goto/DotNetSDKV3/bedrock-runtime-2023-09-30/ConverseStream)sezione *AWS SDK per .NET API Reference*.
------
#### [ Java ]
**SDK per Java 2.x**
C'è di più su GitHub. Trova l'esempio completo e scopri di più sulla configurazione e l'esecuzione nel [Repository di esempi di codice AWS](https://github.com/awsdocs/aws-doc-sdk-examples/tree/main/javav2/example_code/bedrock-runtime#code-examples).
Invia un messaggio di testo ad Amazon Nova utilizzando l’API Converse di Bedrock ed elabora il flusso di risposta in tempo reale.
```
import software.amazon.awssdk.auth.credentials.DefaultCredentialsProvider;
import software.amazon.awssdk.regions.Region;
import software.amazon.awssdk.services.bedrockruntime.BedrockRuntimeAsyncClient;
import software.amazon.awssdk.services.bedrockruntime.model.*;
import java.util.concurrent.ExecutionException;
/**
* This example demonstrates how to use the Amazon Nova foundation models with an
* asynchronous Amazon Bedrock runtime client to generate streaming text responses.
* It shows how to:
* - Set up the Amazon Bedrock runtime client
* - Create a message
* - Configure a streaming request
* - Set up a stream handler to process the response chunks
* - Process the streaming response
*/
public class ConverseStream {
public static void converseStream() {
// Step 1: Create the Amazon Bedrock runtime client
// The runtime client handles the communication with AI models on Amazon Bedrock
BedrockRuntimeAsyncClient client = BedrockRuntimeAsyncClient.builder()
.credentialsProvider(DefaultCredentialsProvider.create())
.region(Region.US_EAST_1)
.build();
// Step 2: Specify which model to use
// Available Amazon Nova models and their characteristics:
// - Amazon Nova Micro: Text-only model optimized for lowest latency and cost
// - Amazon Nova Lite: Fast, low-cost multimodal model for image, video, and text
// - Amazon Nova Pro: Advanced multimodal model balancing accuracy, speed, and cost
//
// For the latest available models, see:
// https://docs.aws.amazon.com/bedrock/latest/userguide/models-supported.html
String modelId = "amazon.nova-lite-v1:0";
// Step 3: Create the message
// The message includes the text prompt and specifies that it comes from the user
var inputText = "Describe the purpose of a 'hello world' program in one paragraph";
var message = Message.builder()
.content(ContentBlock.fromText(inputText))
.role(ConversationRole.USER)
.build();
// Step 4: Configure the request
// Optional parameters to control the model's response:
// - maxTokens: maximum number of tokens to generate
// - temperature: randomness (max: 1.0, default: 0.7)
// OR
// - topP: diversity of word choice (max: 1.0, default: 0.9)
// Note: Use either temperature OR topP, but not both
ConverseStreamRequest request = ConverseStreamRequest.builder()
.modelId(modelId)
.messages(message)
.inferenceConfig(config -> config
.maxTokens(500) // The maximum response length
.temperature(0.5F) // Using temperature for randomness control
//.topP(0.9F) // Alternative: use topP instead of temperature
).build();
// Step 5: Set up the stream handler
// The stream handler processes chunks of the response as they arrive
// - onContentBlockDelta: Processes each text chunk
// - onError: Handles any errors during streaming
var streamHandler = ConverseStreamResponseHandler.builder()
.subscriber(ConverseStreamResponseHandler.Visitor.builder()
.onContentBlockDelta(chunk -> {
System.out.print(chunk.delta().text());
System.out.flush(); // Ensure immediate output of each chunk
}).build())
.onError(err -> System.err.printf("Can't invoke '%s': %s", modelId, err.getMessage()))
.build();
// Step 6: Send the streaming request and process the response
// - Send the request to the model
// - Attach the handler to process response chunks as they arrive
// - Handle any errors during streaming
try {
client.converseStream(request, streamHandler).get();
} catch (ExecutionException | InterruptedException e) {
System.err.printf("Can't invoke '%s': %s", modelId, e.getCause().getMessage());
}
}
public static void main(String[] args) {
converseStream();
}
}
```
+ Per i dettagli sull'API, consulta la [ConverseStream](https://docs.aws.amazon.com/goto/SdkForJavaV2/bedrock-runtime-2023-09-30/ConverseStream)sezione *AWS SDK for Java 2.x API Reference*.
------
#### [ JavaScript ]
**SDK per JavaScript (v3)**
C'è altro da fare. GitHub Trova l'esempio completo e scopri di più sulla configurazione e l'esecuzione nel [Repository di esempi di codice AWS](https://github.com/awsdocs/aws-doc-sdk-examples/tree/main/javascriptv3/example_code/bedrock-runtime#code-examples).
Invia un messaggio di testo ad Amazon Nova utilizzando l’API Converse di Bedrock ed elabora il flusso di risposta in tempo reale.
```
// This example demonstrates how to use the Amazon Nova foundation models
// to generate streaming text responses.
// It shows how to:
// - Set up the Amazon Bedrock runtime client
// - Create a message
// - Configure a streaming request
// - Process the streaming response
import {
BedrockRuntimeClient,
ConversationRole,
ConverseStreamCommand,
} from "@aws-sdk/client-bedrock-runtime";
// Step 1: Create the Amazon Bedrock runtime client
// Credentials will be automatically loaded from the environment
const client = new BedrockRuntimeClient({ region: "us-east-1" });
// Step 2: Specify which model to use
// Available Amazon Nova models and their characteristics:
// - Amazon Nova Micro: Text-only model optimized for lowest latency and cost
// - Amazon Nova Lite: Fast, low-cost multimodal model for image, video, and text
// - Amazon Nova Pro: Advanced multimodal model balancing accuracy, speed, and cost
//
// For the most current model IDs, see:
// https://docs.aws.amazon.com/bedrock/latest/userguide/models-supported.html
const modelId = "amazon.nova-lite-v1:0";
// Step 3: Create the message
// The message includes the text prompt and specifies that it comes from the user
const inputText =
"Describe the purpose of a 'hello world' program in one paragraph";
const message = {
content: [{ text: inputText }],
role: ConversationRole.USER,
};
// Step 4: Configure the streaming request
// Optional parameters to control the model's response:
// - maxTokens: maximum number of tokens to generate
// - temperature: randomness (max: 1.0, default: 0.7)
// OR
// - topP: diversity of word choice (max: 1.0, default: 0.9)
// Note: Use either temperature OR topP, but not both
const request = {
modelId,
messages: [message],
inferenceConfig: {
maxTokens: 500, // The maximum response length
temperature: 0.5, // Using temperature for randomness control
//topP: 0.9, // Alternative: use topP instead of temperature
},
};
// Step 5: Send and process the streaming request
// - Send the request to the model
// - Process each chunk of the streaming response
try {
const response = await client.send(new ConverseStreamCommand(request));
for await (const chunk of response.stream) {
if (chunk.contentBlockDelta) {
// Print each text chunk as it arrives
process.stdout.write(chunk.contentBlockDelta.delta?.text || "");
}
}
} catch (error) {
console.error(`ERROR: Can't invoke '${modelId}'. Reason: ${error.message}`);
process.exitCode = 1;
}
```
+ Per i dettagli sull'API, consulta la [ConverseStream](https://docs.aws.amazon.com/AWSJavaScriptSDK/v3/latest/client/bedrock-runtime/command/ConverseStreamCommand)sezione *AWS SDK per JavaScript API Reference*.
------
#### [ Kotlin ]
**SDK per Kotlin**
C'è di più su GitHub. Trova l'esempio completo e scopri di più sulla configurazione e l'esecuzione nel [Repository di esempi di codice AWS](https://github.com/awsdocs/aws-doc-sdk-examples/tree/main/kotlin/services/bedrock-runtime#code-examples).
Invia un messaggio di testo ad Amazon Nova utilizzando l’API Converse di Bedrock ed elabora il flusso di risposta in tempo reale.
```
import aws.sdk.kotlin.services.bedrockruntime.BedrockRuntimeClient
import aws.sdk.kotlin.services.bedrockruntime.model.ContentBlock
import aws.sdk.kotlin.services.bedrockruntime.model.ConversationRole
import aws.sdk.kotlin.services.bedrockruntime.model.ConverseStreamOutput
import aws.sdk.kotlin.services.bedrockruntime.model.ConverseStreamRequest
import aws.sdk.kotlin.services.bedrockruntime.model.Message
/**
* This example demonstrates how to use the Amazon Nova foundation models
* to generate streaming text responses.
* It shows how to:
* - Set up the Amazon Bedrock runtime client
* - Create a message with a prompt
* - Configure a streaming request with parameters
* - Process the response stream in real time
*/
suspend fun main() {
converseStream()
}
suspend fun converseStream(): String {
// A buffer to collect the complete response
val completeResponseBuffer = StringBuilder()
// Create and configure the Bedrock runtime client
BedrockRuntimeClient { region = "us-east-1" }.use { client ->
// Specify the model ID. For the latest available models, see:
// https://docs.aws.amazon.com/bedrock/latest/userguide/models-supported.html
val modelId = "amazon.nova-lite-v1:0"
// Create the message with the user's prompt
val prompt = "Describe the purpose of a 'hello world' program in a paragraph."
val message = Message {
role = ConversationRole.User
content = listOf(ContentBlock.Text(prompt))
}
// Configure the request with optional model parameters
val request = ConverseStreamRequest {
this.modelId = modelId
messages = listOf(message)
inferenceConfig {
maxTokens = 500 // Maximum response length
temperature = 0.5F // Lower values: more focused output
// topP = 0.8F // Alternative to temperature
}
}
// Process the streaming response
runCatching {
client.converseStream(request) { response ->
response.stream?.collect { chunk ->
when (chunk) {
is ConverseStreamOutput.ContentBlockDelta -> {
// Process each text chunk as it arrives
chunk.value.delta?.asText()?.let { text ->
print(text)
System.out.flush() // Ensure immediate output
completeResponseBuffer.append(text)
}
}
else -> {} // Other output block types can be handled as needed
}
}
}
}.onFailure { error ->
error.message?.let { e -> System.err.println("ERROR: Can't invoke '$modelId'. Reason: $e") }
throw RuntimeException("Failed to generate text with model $modelId: $error", error)
}
}
return completeResponseBuffer.toString()
}
```
+ Per i dettagli sull'API, [ConverseStream](https://sdk.amazonaws.com/kotlin/api/latest/index.html)consulta *AWS SDK for Kotlin* API reference.
------
#### [ Python ]
**SDK per Python (Boto3)**
C'è altro su. GitHub Trova l'esempio completo e scopri di più sulla configurazione e l'esecuzione nel [Repository di esempi di codice AWS](https://github.com/awsdocs/aws-doc-sdk-examples/tree/main/python/example_code/bedrock-runtime#code-examples).
Invia un messaggio di testo ad Amazon Nova utilizzando l’API Converse di Bedrock ed elabora il flusso di risposta in tempo reale.
```
# Use the Conversation API to send a text message to Amazon Nova Text
# and print the response stream.
import boto3
from botocore.exceptions import ClientError
# Create a Bedrock Runtime client in the AWS Region you want to use.
client = boto3.client("bedrock-runtime", region_name="us-east-1")
# Set the model ID, e.g., Amazon Nova Lite.
model_id = "amazon.nova-lite-v1:0"
# Start a conversation with the user message.
user_message = "Describe the purpose of a 'hello world' program in one line."
conversation = [
{
"role": "user",
"content": [{"text": user_message}],
}
]
try:
# Send the message to the model, using a basic inference configuration.
streaming_response = client.converse_stream(
modelId=model_id,
messages=conversation,
inferenceConfig={"maxTokens": 512, "temperature": 0.5, "topP": 0.9},
)
# Extract and print the streamed response text in real-time.
for chunk in streaming_response["stream"]:
if "contentBlockDelta" in chunk:
text = chunk["contentBlockDelta"]["delta"]["text"]
print(text, end="")
except (ClientError, Exception) as e:
print(f"ERROR: Can't invoke '{model_id}'. Reason: {e}")
exit(1)
```
+ Per i dettagli sull'API, consulta [ConverseStream AWS](https://docs.aws.amazon.com/goto/boto3/bedrock-runtime-2023-09-30/ConverseStream)*SDK for Python (Boto3) API Reference*.
------
#### [ Swift ]
**SDK per Swift**
C'è altro da fare. GitHub Trova l'esempio completo e scopri di più sulla configurazione e l'esecuzione nel [Repository di esempi di codice AWS](https://github.com/awsdocs/aws-doc-sdk-examples/tree/main/swift/example_code/bedrock-runtime#code-examples).
Invia un messaggio di testo ad Amazon Nova utilizzando l’API Converse di Bedrock ed elabora il flusso di risposta in tempo reale.
```
// An example demonstrating how to use the Conversation API to send a text message
// to Amazon Nova and print the response stream
import AWSBedrockRuntime
func printConverseStream(_ textPrompt: String) async throws {
// Create a Bedrock Runtime client in the AWS Region you want to use.
let config =
try await BedrockRuntimeClient.BedrockRuntimeClientConfiguration(
region: "us-east-1"
)
let client = BedrockRuntimeClient(config: config)
// Set the model ID.
let modelId = "amazon.nova-lite-v1:0"
// Start a conversation with the user message.
let message = BedrockRuntimeClientTypes.Message(
content: [.text(textPrompt)],
role: .user
)
// Optionally use inference parameters.
let inferenceConfig =
BedrockRuntimeClientTypes.InferenceConfiguration(
maxTokens: 512,
stopSequences: ["END"],
temperature: 0.5,
topp: 0.9
)
// Create the ConverseStreamInput to send to the model.
let input = ConverseStreamInput(
inferenceConfig: inferenceConfig, messages: [message], modelId: modelId)
// Send the ConverseStreamInput to the model.
let response = try await client.converseStream(input: input)
// Extract the streaming response.
guard let stream = response.stream else {
print("No stream available")
return
}
// Extract and print the streamed response text in real-time.
for try await event in stream {
switch event {
case .messagestart(_):
print("\nNova Lite:")
case .contentblockdelta(let deltaEvent):
if case .text(let text) = deltaEvent.delta {
print(text, terminator: "")
}
default:
break
}
}
}
```
+ Per i dettagli sull'API, consulta la [ConverseStream](https://sdk.amazonaws.com/swift/api/awsbedrockruntime/latest/documentation/awsbedrockruntime/bedrockruntimeclient/conversestream(input:))guida di riferimento all'*API AWS SDK for Swift*.
------
# Generare un’immagine
Gli esempi di codice seguenti mostrano come invocare Amazon Nova Canvas in Amazon Bedrock per generare un’immagine.
------
#### [ .NET ]
**SDK per .NET**
C'è altro su. GitHub Trova l'esempio completo e scopri di più sulla configurazione e l'esecuzione nel [Repository di esempi di codice AWS](https://github.com/awsdocs/aws-doc-sdk-examples/tree/main/dotnetv3/Bedrock-runtime#code-examples).
Crea un’immagine con Amazon Nova Canvas.
```
// Use the native inference API to create an image with Amazon Nova Canvas.
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.
var modelId = "amazon.nova-canvas-v1:0";
// Define the image generation prompt for the model.
var prompt = "A stylized picture of a cute old steampunk robot.";
// Create a random seed between 0 and 858,993,459
int seed = new Random().Next(0, 858993460);
//Format the request payload using the model's native structure.
var nativeRequest = JsonSerializer.Serialize(new
{
taskType = "TEXT_IMAGE",
textToImageParams = new
{
text = prompt
},
imageGenerationConfig = new
{
seed,
quality = "standard",
width = 512,
height = 512,
numberOfImages = 1
}
});
// 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 the image data.
var base64Image = modelResponse["images"]?[0].ToString() ?? "";
// Save the image in a local folder
string savedPath = AmazonNovaCanvas.InvokeModel.SaveBase64Image(base64Image);
Console.WriteLine($"Image saved to: {savedPath}");
}
catch (AmazonBedrockRuntimeException e)
{
Console.WriteLine($"ERROR: Can't invoke '{modelId}'. Reason: {e.Message}");
throw;
}
```
+ Per i dettagli sull'API, consulta la [InvokeModel](https://docs.aws.amazon.com/goto/DotNetSDKV3/bedrock-runtime-2023-09-30/InvokeModel)sezione *AWS SDK per .NET API Reference*.
------
#### [ Java ]
**SDK per Java 2.x**
C'è di più su GitHub. Trova l'esempio completo e scopri di più sulla configurazione e l'esecuzione nel [Repository di esempi di codice AWS](https://github.com/awsdocs/aws-doc-sdk-examples/tree/main/javav2/example_code/bedrock-runtime#code-examples).
Crea un’immagine con Amazon Nova Canvas.
```
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;
import software.amazon.awssdk.services.bedrockruntime.model.InvokeModelResponse;
import java.security.SecureRandom;
import java.util.Base64;
import static com.example.bedrockruntime.libs.ImageTools.displayImage;
/**
* This example demonstrates how to use Amazon Nova Canvas to generate images.
* It shows how to:
* - Set up the Amazon Bedrock runtime client
* - Configure the image generation parameters
* - Send a request to generate an image
* - Process the response and handle the generated image
*/
public class InvokeModel {
public static byte[] invokeModel() {
// Step 1: Create the Amazon Bedrock runtime client
// The runtime client handles the communication with AI models on Amazon Bedrock
BedrockRuntimeClient client = BedrockRuntimeClient.builder()
.credentialsProvider(DefaultCredentialsProvider.create())
.region(Region.US_EAST_1)
.build();
// Step 2: Specify which model to use
// For the latest available models, see:
// https://docs.aws.amazon.com/bedrock/latest/userguide/models-supported.html
String modelId = "amazon.nova-canvas-v1:0";
// Step 3: Configure the generation parameters and create the request
// First, set the main parameters:
// - prompt: Text description of the image to generate
// - seed: Random number for reproducible generation (0 to 858,993,459)
String prompt = "A stylized picture of a cute old steampunk robot";
int seed = new SecureRandom().nextInt(858_993_460);
// Then, create the request using a template with the following structure:
// - taskType: TEXT_IMAGE (specifies text-to-image generation)
// - textToImageParams: Contains the text prompt
// - imageGenerationConfig: Contains optional generation settings (seed, quality, etc.)
// For a list of available request parameters, see:
// https://docs.aws.amazon.com/nova/latest/userguide/image-gen-req-resp-structure.html
String request = """
{
"taskType": "TEXT_IMAGE",
"textToImageParams": {
"text": "{{prompt}}"
},
"imageGenerationConfig": {
"seed": {{seed}},
"quality": "standard"
}
}"""
.replace("{{prompt}}", prompt)
.replace("{{seed}}", String.valueOf(seed));
// Step 4: Send and process the request
// - Send the request to the model using InvokeModelResponse
// - Extract the Base64-encoded image from the JSON response
// - Convert the encoded image to a byte array and return it
try {
InvokeModelResponse response = client.invokeModel(builder -> builder
.modelId(modelId)
.body(SdkBytes.fromUtf8String(request))
);
JSONObject responseBody = new JSONObject(response.body().asUtf8String());
// Convert the Base64 string to byte array for better handling
return Base64.getDecoder().decode(
new JSONPointer("/images/0").queryFrom(responseBody).toString()
);
} catch (SdkClientException e) {
System.err.printf("ERROR: Can't invoke '%s'. Reason: %s%n", modelId, e.getMessage());
throw new RuntimeException(e);
}
}
public static void main(String[] args) {
System.out.println("Generating image. This may take a few seconds...");
byte[] imageData = invokeModel();
displayImage(imageData);
}
}
```
+ Per i dettagli sull'API, consulta la [InvokeModel](https://docs.aws.amazon.com/goto/SdkForJavaV2/bedrock-runtime-2023-09-30/InvokeModel)sezione *AWS SDK for Java 2.x API Reference*.
------
#### [ JavaScript ]
**SDK per JavaScript (v3)**
C'è altro da fare. GitHub Trova l'esempio completo e scopri di più sulla configurazione e l'esecuzione nel [Repository di esempi di codice AWS](https://github.com/awsdocs/aws-doc-sdk-examples/tree/main/javascriptv3/example_code/bedrock-runtime#code-examples).
Crea un’immagine con Amazon Nova Canvas.
```
import {
BedrockRuntimeClient,
InvokeModelCommand,
} from "@aws-sdk/client-bedrock-runtime";
import { saveImage } from "../../utils/image-creation.js";
import { fileURLToPath } from "node:url";
/**
* This example demonstrates how to use Amazon Nova Canvas to generate images.
* It shows how to:
* - Set up the Amazon Bedrock runtime client
* - Configure the image generation parameters
* - Send a request to generate an image
* - Process the response and handle the generated image
*
* @returns {Promise} Base64-encoded image data
*/
export const invokeModel = async () => {
// Step 1: Create the Amazon Bedrock runtime client
// Credentials will be automatically loaded from the environment
const client = new BedrockRuntimeClient({ region: "us-east-1" });
// Step 2: Specify which model to use
// For the latest available models, see:
// https://docs.aws.amazon.com/bedrock/latest/userguide/models-supported.html
const modelId = "amazon.nova-canvas-v1:0";
// Step 3: Configure the request payload
// First, set the main parameters:
// - prompt: Text description of the image to generate
// - seed: Random number for reproducible generation (0 to 858,993,459)
const prompt = "A stylized picture of a cute old steampunk robot";
const seed = Math.floor(Math.random() * 858993460);
// Then, create the payload using the following structure:
// - taskType: TEXT_IMAGE (specifies text-to-image generation)
// - textToImageParams: Contains the text prompt
// - imageGenerationConfig: Contains optional generation settings (seed, quality, etc.)
// For a list of available request parameters, see:
// https://docs.aws.amazon.com/nova/latest/userguide/image-gen-req-resp-structure.html
const payload = {
taskType: "TEXT_IMAGE",
textToImageParams: {
text: prompt,
},
imageGenerationConfig: {
seed,
quality: "standard",
},
};
// Step 4: Send and process the request
// - Embed the payload in a request object
// - Send the request to the model
// - Extract and return the generated image data from the response
try {
const request = {
modelId,
body: JSON.stringify(payload),
};
const response = await client.send(new InvokeModelCommand(request));
const decodedResponseBody = new TextDecoder().decode(response.body);
// The response includes an array of base64-encoded PNG images
/** @type {{images: string[]}} */
const responseBody = JSON.parse(decodedResponseBody);
return responseBody.images[0]; // Base64-encoded image data
} catch (error) {
console.error(`ERROR: Can't invoke '${modelId}'. Reason: ${error.message}`);
throw error;
}
};
// If run directly, execute the example and save the generated image
if (process.argv[1] === fileURLToPath(import.meta.url)) {
console.log("Generating image. This may take a few seconds...");
invokeModel()
.then(async (imageData) => {
const imagePath = await saveImage(imageData, "nova-canvas");
// Example path: javascriptv3/example_code/bedrock-runtime/output/nova-canvas/image-01.png
console.log(`Image saved to: ${imagePath}`);
})
.catch((error) => {
console.error("Execution failed:", error);
process.exitCode = 1;
});
}
```
+ Per i dettagli sull'API, consulta la [InvokeModel](https://docs.aws.amazon.com/AWSJavaScriptSDK/v3/latest/client/bedrock-runtime/command/InvokeModelCommand)sezione *AWS SDK per JavaScript API Reference*.
------
#### [ Python ]
**SDK per Python (Boto3)**
C'è di più su GitHub. Trova l'esempio completo e scopri di più sulla configurazione e l'esecuzione nel [Repository di esempi di codice AWS](https://github.com/awsdocs/aws-doc-sdk-examples/tree/main/python/example_code/bedrock-runtime#code-examples).
Crea un’immagine con Amazon Nova Canvas.
```
# Use the native inference API to create an image with Amazon Nova Canvas
import base64
import json
import os
import random
import boto3
# 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.
model_id = "amazon.nova-canvas-v1:0"
# Define the image generation prompt for the model.
prompt = "A stylized picture of a cute old steampunk robot."
# Generate a random seed between 0 and 858,993,459
seed = random.randint(0, 858993460)
# Format the request payload using the model's native structure.
native_request = {
"taskType": "TEXT_IMAGE",
"textToImageParams": {"text": prompt},
"imageGenerationConfig": {
"seed": seed,
"quality": "standard",
"height": 512,
"width": 512,
"numberOfImages": 1,
},
}
# Convert the native request to JSON.
request = json.dumps(native_request)
# Invoke the model with the request.
response = client.invoke_model(modelId=model_id, body=request)
# Decode the response body.
model_response = json.loads(response["body"].read())
# Extract the image data.
base64_image_data = model_response["images"][0]
# Save the generated image to a local folder.
i, output_dir = 1, "output"
if not os.path.exists(output_dir):
os.makedirs(output_dir)
while os.path.exists(os.path.join(output_dir, f"nova_canvas_{i}.png")):
i += 1
image_data = base64.b64decode(base64_image_data)
image_path = os.path.join(output_dir, f"nova_canvas_{i}.png")
with open(image_path, "wb") as file:
file.write(image_data)
print(f"The generated image has been saved to {image_path}")
```
+ Per i dettagli sull'API, consulta [InvokeModel AWS](https://docs.aws.amazon.com/goto/boto3/bedrock-runtime-2023-09-30/InvokeModel)*SDK for Python (Boto3) API Reference*.
------
#### [ Swift ]
**SDK per Swift**
C'è altro da fare. GitHub Trova l'esempio completo e scopri di più sulla configurazione e l'esecuzione nel [Repository di esempi di codice AWS](https://github.com/awsdocs/aws-doc-sdk-examples/tree/main/swift/example_code/bedrock-runtime#code-examples).
Crea un’immagine con Amazon Nova Canvas.
```
// Use the native inference API to create an image with Amazon Nova Canvas
import AWSBedrockRuntime
import AWSSDKIdentity
import Foundation
struct NovaImageOutput: Decodable {
let images: [Data]
}
func generateImage(_ textPrompt: String) async throws {
// Create a Bedrock Runtime client in the AWS Region you want to use.
let config =
try await BedrockRuntimeClient.BedrockRuntimeClientConfiguration(
region: "us-east-1"
)
config.awsCredentialIdentityResolver = try SSOAWSCredentialIdentityResolver()
let client = BedrockRuntimeClient(config: config)
// Set the model ID.
let modelId = "amazon.nova-canvas-v1:0"
// Format the request payload using the model's native structure.
let input = InvokeModelInput(
accept: "application/json",
body: """
{
"textToImageParams": {
"text": "\(textPrompt)"
},
"taskType": "TEXT_IMAGE",
"imageGenerationConfig": {
"seed": 42,
"quality": "standard",
"width": 512,
"height": 512,
"numberOfImages": 1
}
}
""".data(using: .utf8),
modelId: modelId
)
// Invoke the model with the request.
let response = try await client.invokeModel(input: input)
// Decode the response body.
let output = try JSONDecoder().decode(NovaImageOutput.self, from: response.body!)
// Extract the image data.
guard let data = output.images.first else {
print("No image data found")
return
}
// Save the generated image to a local folder.
let fileURL = URL.documentsDirectory.appending(path: "nova_canvas.png")
print(fileURL)
try data.write(to: fileURL)
print("Image is saved at \(fileURL)")
}
```
+ Per i dettagli sull'API, consulta la [InvokeModel](https://sdk.amazonaws.com/swift/api/awsbedrockruntime/latest/documentation/awsbedrockruntime/bedrockruntimeclient/invokemodel(input:))guida di riferimento all'*API AWS SDK for Swift*.
------
# Generare un video
Gli esempi di codice seguenti mostrano come utilizzare Amazon Nova Reel per generare un video da un prompt di testo.
------
#### [ Python ]
**SDK per Python (Boto3)**
C'è altro su. GitHub Trova l'esempio completo e scopri di più sulla configurazione e l'esecuzione nel [Repository di esempi di codice AWS](https://github.com/awsdocs/aws-doc-sdk-examples/tree/main/python/example_code/bedrock-runtime#code-examples).
Utilizza Amazon Nova Reel per generare un video da un prompt di testo.
```
"""
This example demonstrates how to use Amazon Nova Reel to generate a video from a text prompt.
It shows how to:
- Set up the Amazon Bedrock runtime client
- Configure a text-to-video request
- Submit an asynchronous job for video generation
- Poll for job completion status
- Access the generated video from S3
"""
import random
import time
import boto3
# Replace with your own S3 bucket to store the generated video
# Format: s3://your-bucket-name
OUTPUT_S3_URI = "s3://REPLACE-WITH-YOUR-S3-BUCKET-NAME"
def start_text_to_video_generation_job(bedrock_runtime, prompt, output_s3_uri):
"""
Starts an asynchronous text-to-video generation job using Amazon Nova Reel.
:param bedrock_runtime: The Bedrock runtime client
:param prompt: The text description of the video to generate
:param output_s3_uri: S3 URI where the generated video will be stored
:return: The invocation ARN of the async job
"""
# Specify the model ID for text-to-video generation
model_id = "amazon.nova-reel-v1:0"
# Generate a random seed between 0 and 2,147,483,646
# This helps ensure unique video generation results
seed = random.randint(0, 2147483646)
# Configure the video generation request with additional parameters
model_input = {
"taskType": "TEXT_VIDEO",
"textToVideoParams": {"text": prompt},
"videoGenerationConfig": {
"fps": 24,
"durationSeconds": 6,
"dimension": "1280x720",
"seed": seed,
},
}
# Specify the S3 location for the output video
output_config = {"s3OutputDataConfig": {"s3Uri": output_s3_uri}}
# Invoke the model asynchronously
response = bedrock_runtime.start_async_invoke(
modelId=model_id, modelInput=model_input, outputDataConfig=output_config
)
invocation_arn = response["invocationArn"]
return invocation_arn
def query_job_status(bedrock_runtime, invocation_arn):
"""
Queries the status of an asynchronous video generation job.
:param bedrock_runtime: The Bedrock runtime client
:param invocation_arn: The ARN of the async invocation to check
:return: The runtime response containing the job status and details
"""
return bedrock_runtime.get_async_invoke(invocationArn=invocation_arn)
def main():
"""
Main function that demonstrates the complete workflow for generating
a video from a text prompt using Amazon Nova Reel.
"""
# Create a Bedrock Runtime client
# Note: Credentials will be loaded from the environment or AWS CLI config
bedrock_runtime = boto3.client("bedrock-runtime", region_name="us-east-1")
# Configure the text prompt and output location
prompt = "Closeup of a cute old steampunk robot. Camera zoom in."
# Verify the S3 URI has been set to a valid bucket
if "REPLACE-WITH-YOUR-S3-BUCKET-NAME" in OUTPUT_S3_URI:
print("ERROR: You must replace the OUTPUT_S3_URI with your own S3 bucket URI")
return
print("Submitting video generation job...")
invocation_arn = start_text_to_video_generation_job(
bedrock_runtime, prompt, OUTPUT_S3_URI
)
print(f"Job started with invocation ARN: {invocation_arn}")
# Poll for job completion
while True:
print("\nPolling job status...")
job = query_job_status(bedrock_runtime, invocation_arn)
status = job["status"]
if status == "Completed":
bucket_uri = job["outputDataConfig"]["s3OutputDataConfig"]["s3Uri"]
print(f"\nSuccess! The video is available at: {bucket_uri}/output.mp4")
break
elif status == "Failed":
print(
f"\nVideo generation failed: {job.get('failureMessage', 'Unknown error')}"
)
break
else:
print("In progress. Waiting 15 seconds...")
time.sleep(15)
if __name__ == "__main__":
main()
```
+ Per informazioni dettagliate sull’API, consulta i seguenti argomenti nella *documentazione di riferimento dell’API AWS SDK per Python (Boto3)*.
+ [GetAsyncInvoke](https://docs.aws.amazon.com/goto/boto3/bedrock-runtime-2023-09-30/GetAsyncInvoke)
+ [StartAsyncInvoke](https://docs.aws.amazon.com/goto/boto3/bedrock-runtime-2023-09-30/StartAsyncInvoke)
------
#### [ Swift ]
**SDK per Swift**
C'è altro da fare. GitHub Trova l'esempio completo e scopri di più sulla configurazione e l'esecuzione nel [Repository di esempi di codice AWS](https://github.com/awsdocs/aws-doc-sdk-examples/tree/main/swift/example_code/bedrock-runtime#code-examples).
Utilizza Amazon Nova Reel per generare un video da un prompt di testo.
```
// This example demonstrates how to use Amazon Nova Reel to generate a video from a text prompt.
// It shows how to:
// - Set up the Amazon Bedrock runtime client
// - Configure a text-to-video request
// - Submit an asynchronous job for video generation
// - Poll for job completion status
// - Access the generated video from S3
import AWSBedrockRuntime
import Foundation
import Smithy
func startTextToVideoGenerationJob(
bedrockRuntimeClient: BedrockRuntimeClient, prompt: String, outputS3Uri: String
) async throws -> String? {
// Specify the model ID for text-to-video generation
let modelId = "amazon.nova-reel-v1:0"
// Configure the video generation request with additional parameters
let modelInputSource: [String: Any] = [
"taskType": "TEXT_VIDEO",
"textToVideoParams": [
"text": "\(prompt)"
],
"videoGenerationConfig": [
"durationSeconds": 6,
"fps": 24,
"dimension": "1280x720",
],
]
let modelInput = try Document.make(from: modelInputSource)
let input = StartAsyncInvokeInput(
modelId: modelId,
modelInput: modelInput,
outputDataConfig: .s3outputdataconfig(
BedrockRuntimeClientTypes.AsyncInvokeS3OutputDataConfig(
s3Uri: outputS3Uri
)
)
)
// Invoke the model asynchronously
let output = try await bedrockRuntimeClient.startAsyncInvoke(input: input)
return output.invocationArn
}
func queryJobStatus(
bedrockRuntimeClient: BedrockRuntimeClient,
invocationArn: String?
) async throws -> GetAsyncInvokeOutput {
try await bedrockRuntimeClient.getAsyncInvoke(
input: GetAsyncInvokeInput(invocationArn: invocationArn))
}
func main() async throws {
// Create a Bedrock Runtime client
let config =
try await BedrockRuntimeClient.BedrockRuntimeClientConfiguration(
region: "us-east-1"
)
let client = BedrockRuntimeClient(config: config)
// Specify the S3 location for the output video
let bucket = "s3://REPLACE-WITH-YOUR-S3-BUCKET-NAM"
print("Submitting video generation job...")
let invocationArn = try await startTextToVideoGenerationJob(
bedrockRuntimeClient: client,
prompt: "A pomegranate juice in a railway station",
outputS3Uri: bucket
)
print("Job started with invocation ARN: \(String(describing:invocationArn))")
// Poll for job completion
var status: BedrockRuntimeClientTypes.AsyncInvokeStatus?
var isReady = false
var hasFailed = false
while !isReady && !hasFailed {
print("\nPolling job status...")
status = try await queryJobStatus(
bedrockRuntimeClient: client, invocationArn: invocationArn
).status
switch status {
case .completed:
isReady = true
print("Video is ready\nCheck S3 bucket: \(bucket)")
case .failed:
hasFailed = true
print("Something went wrong")
case .inProgress:
print("Job is in progress...")
try await Task.sleep(nanoseconds: 15 * 1_000_000_000) // 15 seconds
default:
isReady = true
}
}
}
do {
try await main()
} catch {
print("An error occurred: \(error)")
}
```
+ Per informazioni dettagliate sull’API, consulta i seguenti argomenti nella *documentazione di riferimento dell’API SDK AWS per Swift*.
+ [GetAsyncInvoke](https://sdk.amazonaws.com/swift/api/awsbedrockruntime/latest/documentation/awsbedrockruntime/bedrockruntimeclient/getasyncinvoke(input:))
+ [StartAsyncInvoke](https://sdk.amazonaws.com/swift/api/awsbedrockruntime/latest/documentation/awsbedrockruntime/bedrockruntimeclient/startasyncinvoke(input:))
------
# Utilizzare uno strumento con Amazon Nova
I seguenti esempi di codice mostrano come creare un'interazione tipica tra un'applicazione, un modello di intelligenza artificiale generativa e strumenti connessi o come APIs mediare le interazioni tra l'IA e il mondo esterno. Viene utilizzato l’esempio di collegamento di un’API esterna per il meteo al modello di IA in modo che vengano fornite informazioni meteorologiche in tempo reale in base sull’input dell’utente.
------
#### [ .NET ]
**SDK per .NET**
C'è altro da fare. GitHub Trova l'esempio completo e scopri di più sulla configurazione e l'esecuzione nel [Repository di esempi di codice AWS](https://github.com/awsdocs/aws-doc-sdk-examples/tree/main/dotnetv3/Bedrock-runtime/Scenarios/ConverseToolScenario#code-examples).
Esecuzione principale del flusso dello scenario. Questo scenario orchestra la conversazione tra l’utente, l’API Converse per Amazon Bedrock e uno strumento meteo.
```
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("Microsoft", LogLevel.Trace))
.ConfigureServices((_, services) =>
services.AddHttpClient()
.AddSingleton(_ => new AmazonBedrockRuntimeClient(RegionEndpoint.USEast1)) // Specify a region that has access to the chosen model.
.AddTransient()
.AddTransient()
.RemoveAll()
)
.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));
}
}
///
/// Populate the services for use within the console application.
///
/// The services host.
private static void ServicesSetup(IHost host)
{
_bedrockActionsWrapper = host.Services.GetRequiredService();
_weatherTool = host.Services.GetRequiredService();
}
///
/// Starts the conversation with the user and handles the interaction with Bedrock.
///
/// The conversation array.
public static async Task> RunConversationAsync()
{
// Print the greeting and a short user guide
PrintHeader();
// Start with an empty conversation
var conversation = new List();
// 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 { 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;
}
///
/// Sends the conversation, the system prompt, and the tool spec to Amazon Bedrock, and returns the response.
///
/// The conversation history including the next message to send.
/// The response from Amazon Bedrock.
private static async Task SendConversationToBedrock(List 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());
}
///
/// Processes the response received via Amazon Bedrock and performs the necessary actions based on the stop reason.
///
/// The model's response returned via Amazon Bedrock.
/// The conversation history.
/// The maximum number of recursive calls allowed.
private static async Task ProcessModelResponseAsync(ConverseResponse modelResponse, List 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";
}
}
}
///
/// 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.
///
/// The model's response containing the tool use request.
/// The conversation history.
/// The maximum number of recursive calls allowed.
public static async Task HandleToolUseAsync(ConverseOutput modelResponse, List conversation, int maxRecursion)
{
// Initialize an empty list of tool results
var toolResults = new List();
// 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()
{ 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);
}
///
/// 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.
///
/// The payload containing the tool name and input data.
/// The tool's response or an error message.
public static async Task 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 = $"\tThe requested tool with name '{toolName}' does not exist.";
return new ToolResponse { ToolUseId = payload.ToolUseId, Content = new { error = true, message = errorMessage } };
}
}
///
/// Prompts the user for input and returns the user's response.
/// Returns null if the user enters 'x' to exit.
///
/// The prompt to display to the user.
/// The user's input or null if the user chooses to exit.
private static async Task GetUserInputAsync(string prompt = "\tYour weather info request:")
{
var userInput = default_prompt;
if (_interactive)
{
Console.WriteLine(new string('*', 80));
Console.WriteLine($"{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;
}
///
/// Logs the welcome message and usage guide for the tool use demo.
///
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!
");
}
///
/// Logs the footer information for the tool use demo.
///
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
=================================================
");
}
///
/// Logs information about the tool use.
///
/// The name of the tool being used.
/// The input data for the tool.
public static void PrintToolUse(string toolName, Dictionary inputData)
{
Console.WriteLine($"\n\tInvoking tool: {toolName} with input: {inputData["latitude"].ToString()}, {inputData["longitude"].ToString()}...\n");
}
///
/// Logs the model's response.
///
/// The model's response message.
public static void PrintModelResponse(string message)
{
Console.WriteLine("\tThe model's response:\n");
Console.WriteLine(message);
Console.WriteLine();
}
}
```
Strumento meteo utilizzato dalla demo. Questo file definisce le specifiche dello strumento e implementa la logica per recuperare i dati meteorologici utilizzando l’API Open-Meteo.
```
using Amazon.BedrockRuntime.Model;
using Amazon.Runtime.Documents;
using Microsoft.Extensions.Logging;
namespace ConverseToolScenario;
///
/// Weather tool that will be invoked when requested by the Bedrock response.
///
public class WeatherTool
{
private readonly ILogger _logger;
private readonly IHttpClientFactory _httpClientFactory;
public WeatherTool(ILogger logger, IHttpClientFactory httpClientFactory)
{
_logger = logger;
_httpClientFactory = httpClientFactory;
}
///
/// 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.
///
/// The tool specification for the Weather tool.
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;
}
///
/// 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.
///
/// The latitude of the location.
/// The longitude of the location.
/// The weather data or an error message.
public async Task 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}¤t_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;
}
}
}
```
Azione dell’API Converse con una configurazione dello strumento.
```
///
/// Wrapper class for interacting with the Amazon Bedrock Converse API.
///
public class BedrockActionsWrapper
{
private readonly IAmazonBedrockRuntime _bedrockClient;
private readonly ILogger _logger;
///
/// Initializes a new instance of the class.
///
/// The Bedrock Converse API client.
/// The logger instance.
public BedrockActionsWrapper(IAmazonBedrockRuntime bedrockClient, ILogger logger)
{
_bedrockClient = bedrockClient;
_logger = logger;
}
///
/// Sends a Converse request to the Amazon Bedrock Converse API.
///
/// The Bedrock Model Id.
/// A system prompt instruction.
/// The array of messages in the conversation.
/// The specification for a tool.
/// The response of the model.
public async Task SendConverseRequestAsync(string modelId, string systemPrompt, List conversation, ToolSpecification toolSpec)
{
try
{
var request = new ConverseRequest()
{
ModelId = modelId,
System = new List()
{
new SystemContentBlock()
{
Text = systemPrompt
}
},
Messages = conversation,
ToolConfig = new ToolConfiguration()
{
Tools = new List()
{
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;
}
}
}
```
+ Per informazioni dettagliate sull’API, consulta [Converse](https://docs.aws.amazon.com/goto/DotNetSDKV3/bedrock-runtime-2023-09-30/Converse) nella *documentazione di riferimento dell’API AWS SDK per .NET *.
------
#### [ Java ]
**SDK per Java 2.x**
C'è dell'altro GitHub. Trova l'esempio completo e scopri di più sulla configurazione e l'esecuzione nel [Repository di esempi di codice AWS](https://github.com/awsdocs/aws-doc-sdk-examples/tree/main/javav2/example_code/bedrock-runtime#code-examples).
Esecuzione principale del flusso dello scenario. Questo scenario orchestra la conversazione tra l’utente, l’API Converse per Amazon Bedrock e uno strumento meteo.
```
/*
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 runConversation(Scanner scanner) {
List 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 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 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 conversation, int maxRecursion) {
List 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 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 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 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 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 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 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
=================================================
""");
}
}
```
Strumento meteo utilizzato dalla demo. Questo file definisce le specifiche dello strumento e implementa la logica per recuperare i dati meteorologici utilizzando l’API Open-Meteo.
```
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 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 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 propertiesMap = new HashMap<>();
propertiesMap.put("latitude", Document.fromMap(latitudeMap));
propertiesMap.put("longitude", Document.fromMap(longitudeMap));
// Create the "required" array
List requiredList = new ArrayList<>();
requiredList.add(Document.fromString("latitude"));
requiredList.add(Document.fromString("longitude"));
// Create the root object
Map 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¤t_weather=True", endpoint, latitude, longitude);
HttpRequest request = HttpRequest.newBuilder()
.uri(URI.create(url))
.build();
try {
HttpResponse response = httpClient.send(request, HttpResponse.BodyHandlers.ofString());
if (response.statusCode() == 200) {
String weatherJson = response.body();
System.out.println(weatherJson);
ObjectMapper objectMapper = new ObjectMapper();
Map rawMap = objectMapper.readValue(weatherJson, new TypeReference>() {});
Map 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 convertToDocumentMap(Map inputMap) {
Map result = new HashMap<>();
for (Map.Entry 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) 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
}
}
```
Azione dell’API Converse con una configurazione dello strumento.
```
/**
* 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 conversation, ToolSpecification toolSpec) {
List 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);
}
}
```
+ Per informazioni dettagliate sull’API, consulta [Converse](https://docs.aws.amazon.com/goto/SdkForJavaV2/bedrock-runtime-2023-09-30/Converse) nella *documentazione di riferimento dell’API AWS SDK for Java 2.x *.
------
#### [ JavaScript ]
**SDK per JavaScript (v3)**
C'è altro da fare. GitHub Trova l'esempio completo e scopri di più sulla configurazione e l'esecuzione nel [Repository di esempi di codice AWS](https://github.com/awsdocs/aws-doc-sdk-examples/tree/main/javascriptv3/example_code/bedrock-runtime/scenarios/converse_tool_scenario#code-examples).
Esecuzione principale del flusso dello scenario. Questo scenario orchestra la conversazione tra l’utente, l’API Converse per Amazon Bedrock e uno strumento meteo.
```
/* 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}¤t_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 });
}
```
+ Per informazioni dettagliate sull’API, consulta [Converse](https://docs.aws.amazon.com/AWSJavaScriptSDK/v3/latest/client/bedrock-runtime/command/ConverseCommand) nella *documentazione di riferimento dell’API AWS SDK per JavaScript *.
------