Universal.GenerativeAI.AmazonBedrock 1.0.0

Universal.GenerativeAI.AmazonBedrock

Implementation of generative AI interfaces using Amazon Bedrock models for text generation and transformation.

Classes

AmazonBedrockTextGenerator

Implements the IAsyncTextGenerator interface to generate text using Amazon Bedrock's AI models.

// Basic usage with default options
var bedrockClient = new AmazonBedrockRuntimeClient();
var generator = new AmazonBedrockTextGenerator(bedrockClient);
string generatedText = generator.Generate("Write a summary of the key features of quantum computing:");

// Asynchronous usage with custom options
var options = new AmazonBedrockTextGenerator.Options
{
    ModelId = "<model ID or inference profile ARN>",
    MaxTokens = 4096
};
var generator = new AmazonBedrockTextGenerator(bedrockClient, options);
string generatedText = await generator.GenerateAsync("Write a summary of the key features of quantum computing:", cancellationToken);

AmazonBedrockTextTransformer

Implements the IAsyncTextTransformer interface to transform text into structured objects using Amazon Bedrock's AI models.

// Basic usage to transform text into a custom type
var bedrockClient = new AmazonBedrockRuntimeClient();
var transformer = new AmazonBedrockTextTransformer(bedrockClient);

// Define a class to hold the structured data
public class Product
{
    public string Name { get; set; }
    public string Description { get; set; }
    public decimal Price { get; set; }
}

// Transform text into a Product instance, with explicit prompt for JSON output
Product product = transformer.Transform<Product>(
    "Extract product details from this text and respond ONLY with valid JSON in the following format:\n" +
    "{\n" +
    "  \"Name\": \"product name\",\n" +
    "  \"Description\": \"product description\",\n" +
    "  \"Price\": numeric price\n" +
    "}\n\n" +
    "Text: Gaming laptop XPS-9000, high-performance gaming with RGB keyboard, priced at $1299.99"
);

// Asynchronous usage with custom options
var options = new AmazonBedrockTextTransformer.Options
{
    ModelId = "anthropic.claude-3-5-sonnet-20240620-v1:0",
    MaxTokens = 2048
};
var transformer = new AmazonBedrockTextTransformer(bedrockClient, options);
List<Product> products = await transformer.TransformAsync<List<Product>>(
    "Extract all products from this text and respond ONLY with valid JSON as an array of objects in the following format:\n" +
    "[\n" +
    "  {\n" +
    "    \"Name\": \"product name\",\n" +
    "    \"Description\": \"product description\",\n" +
    "    \"Price\": numeric price\n" +
    "  },\n" +
    "  ...\n" +
    "]\n\n" +
    "Text: Our latest models include the XPS-9000 gaming laptop ($1299.99) with RGB keyboard, and the UltraSlim X1 ultrabook ($899.50) featuring a 4K touchscreen.",
    cancellationToken
);

Features

• Token Streaming: Automatically handles responses that exceed token limits by streaming content in batches
• Model Selection: Configure different Amazon Bedrock models based on your needs
• Intelligent Type Detection: Automatically determines the appropriate JSON structure based on the target type
• Configurable Token Limits: Adjust batch token limits to optimize for performance or cost