ChatPromptVariantProps

class aws_cdk.aws_bedrock_alpha.ChatPromptVariantProps(*, model, variant_name, prompt_variables=None, messages, inference_configuration=None, system=None, tool_configuration=None)

Bases: CommonPromptVariantProps

(experimental) Properties for creating a chat prompt variant.

Parameters:

• model (IBedrockInvokable) – (experimental) The model which is used to run the prompt. The model could be a foundation model, a custom model, or a provisioned model.

• variant_name (str) – (experimental) The name of the prompt variant.

• prompt_variables (Optional[Sequence[str]]) – (experimental) The variables in the prompt template that can be filled in at runtime. Default: - No variables defined.

• messages (Sequence[ChatMessage]) – (experimental) The messages in the chat prompt. Must include at least one User Message. The messages should alternate between User and Assistant.

• inference_configuration (Optional[PromptInferenceConfiguration]) – (experimental) Inference configuration for the Chat Prompt. Default: - No inference configuration provided.

• system (Optional[str]) – (experimental) Context or instructions for the model to consider before generating a response. Default: - No system message provided.

• tool_configuration (Union[ToolConfiguration, Dict[str, Any], None]) – (experimental) The configuration with available tools to the model and how it must use them. Default: - No tool configuration provided.

Stability:

experimental

ExampleMetadata:

fixture=default infused

Example:

cmk = kms.Key(self, "cmk")

variant_chat = bedrock.PromptVariant.chat(
    variant_name="variant1",
    model=bedrock.BedrockFoundationModel.ANTHROPIC_CLAUDE_3_5_SONNET_V1_0,
    messages=[
        bedrock.ChatMessage.user("From now on, you speak Japanese!"),
        bedrock.ChatMessage.assistant("Konnichiwa!"),
        bedrock.ChatMessage.user("From now on, you speak {{language}}!")
    ],
    system="You are a helpful assistant that only speaks the language you`re told.",
    prompt_variables=["language"],
    tool_configuration=bedrock.ToolConfiguration(
        tool_choice=bedrock.ToolChoice.AUTO,
        tools=[
            bedrock.Tool.function(
                name="top_song",
                description="Get the most popular song played on a radio station.",
                input_schema={
                    "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 WKR."
                        }
                    },
                    "required": ["sign"]
                }
            )
        ]
    )
)

bedrock.Prompt(self, "prompt1",
    prompt_name="prompt-chat",
    description="my first chat prompt",
    default_variant=variant_chat,
    variants=[variant_chat],
    kms_key=cmk
)

Attributes

inference_configuration

(experimental) Inference configuration for the Chat Prompt.

Default:

• No inference configuration provided.

Stability:

experimental

messages

(experimental) The messages in the chat prompt.

Must include at least one User Message. The messages should alternate between User and Assistant.

Stability:

experimental

model

(experimental) The model which is used to run the prompt.

The model could be a foundation model, a custom model, or a provisioned model.

Stability:

experimental

prompt_variables

(experimental) The variables in the prompt template that can be filled in at runtime.

Default:

• No variables defined.

Stability:

experimental

system

(experimental) Context or instructions for the model to consider before generating a response.

Default:

• No system message provided.

Stability:

experimental

tool_configuration

(experimental) The configuration with available tools to the model and how it must use them.

Default:

• No tool configuration provided.

Stability:

experimental

variant_name

(experimental) The name of the prompt variant.

Stability:

experimental