ConversationAgent

pydantic model ConversationAgent

Bases: BaseConversationAgent

LLM agent for conversing about a topic in the target language

Create a new model by parsing and validating input data from keyword arguments.

Raises ValidationError if the input data cannot be parsed to form a valid model.

Fields:

• llm_chain (langchain.chains.llm.LLMChain)

• stop (List | None)

field llm_chain: LLMChain [Required]

Any LLM chain for inference

field stop: List | None = None

LLM end-of-response keyword

greet(inputs: Dict[str, str], callbacks: List[BaseCallbackHandler] | BaseCallbackManager | None = None) → str

Start a new conversation about a topic.

speak(inputs: Dict[str, str], callbacks: List[BaseCallbackHandler] | BaseCallbackManager | None = None) → str

Get single response to user input

Parameters:

• inputs –

  Conversation inputs dictionary. Expects:

  • language: target language (optional)

  • proficiency: user proficiency with target language (optional)

  • topic: the topic to discuss (optional)

  • input: the current user input in the conversation

• callbacks – LangChain LLM callbacks

Returns:

Single LLM response

Return type:

str