Code Reference

The ChatBot Class

class chatbot.chatbot.ChatBot(api_key: str, prompt: str = "You're a nice helpful chatbot.", message_memory: ~chatbot.chatbot.BaseMessageMemory = <chatbot.chatbot.MessageMemory object>, knowledge_base: ~chatbot.chatbot.BaseKnowledgeBase | None = None, gpt_model: str = 'gpt-3.5-turbo')

Bases: object

Wrapper to call OpenAI’s GPT-3.5 API and return a response. Optionally uses child class implementation of BaseMessageMemory and BaseKnowledgeBase classes to configure how memory is stored and how to get context for the user’s query.

Parameters:

  • api_key (str) – The API key for OpenAI’s API

  • prompt (str, optional) – The prompt to use when calling OpenAI’s API, defaults to “You’re a nice helpful chatbot.” You will probably want to change this to make the chatbot more interesting.

  • message_memory (BaseMessageMemory, optional) – The message memory to use, defaults to MessageMemory - which is a simple in-memory implementation.

  • knowledge_base (BaseKnowledgeBase, optional) – The knowledge base to use, defaults to None. If you want to use a knowledge base, you must implement the BaseKnowledgeBase, or use the KnowledgeBaseRedis.

  • gpt_model (str, optional) – The GPT model to use, defaults to “gpt-3.5-turbo-0613”.

get_reply(user_query: str) → str

Get a reply from the chatbot.

Parameters:

user_query (str) – The user’s query

Returns:

The chatbot’s response

Return type:

str

The BaseMessageMemory Class

class chatbot.chatbot.BaseMessageMemory(memory_length: int = 5)

Bases: object

Base class for storing the conversation history. This interface is assumed to be implemented by the ChatBot, so any new child classes must implement these methods.

Parameters:

memory_length (int, optional) – The number of messages to store in the chat history, defaults to 5

add_latest_bot_response()
add_latest_user_query()
get_message_list()
trim_message_list()

The MessageMemory Class

class chatbot.chatbot.MessageMemory(memory_length: int = 5)

Bases: BaseMessageMemory

Stores the conversation history in memory and uses the memory_length parameter to determine how many messages to send to the API to provide context.

Parameters:

memory_length (int, optional) – The number of messages to store in the chat history, defaults to 5

add_latest_bot_response(bot_response: str)

Add the latest bot response to the message queue.

Parameters:

bot_response (str) – The latest bot response

add_latest_user_query(latest_message: str)

Add the latest user query to the message queue.

Parameters:

latest_message (str) – The latest user query

get_message_list() → list[dict]

Returns the message queue. :return: The message queue :rtype: list[dict]

trim_message_list()

Trims the message list to memory_length.

The BaseKnowledgeBase Class

class chatbot.chatbot.BaseKnowledgeBase

Bases: object

Base class for a knowledge base. This interface is assumed to be implemented by the ChatBot, so any new child classes must implement these methods.

get_context(user_query: str)

The KnowledgeBaseRedis Class

class chatbot.chatbot.KnowledgeBaseRedis(redis_url: str, api_key: str)

Bases: BaseKnowledgeBase

A knowledge base that uses Redis to store information and the embeddings for each piece of information. This class uses Redis’ vector search to find the most similar information to the user’s query.

Parameters:

  • redis_url (str) – The URL for the Redis instance

  • api_key (str) – The API key for OpenAI’s API

get_context(user_query: str) → str | None

Get the context for the user’s query.

Parameters:

user_query (str) – The user’s query

Returns:

The context for the user’s query

Return type:

str | None