Skip to content

Latest commit

 

History

History
118 lines (93 loc) · 4.01 KB

chat.md

File metadata and controls

118 lines (93 loc) · 4.01 KB

Chat

Overview

The Chat class in MiniAutoGen is a crucial component designed to manage and store the state of group chats. It provides functionalities for handling messages, maintaining chat context, and managing agents within a chat session.

Class Definition

class Chat

Description:

The Chat class is responsible for managing chat sessions, including message storage, chat context, and agent interactions.

Constructor:

__init__(self, storage_path='groupchat_data', custom_df=None)
  • Purpose: Initializes a Chat object with a specified storage path and an optional custom DataFrame.
  • Parameters:
    • storage_path (str): Path to the storage directory. Defaults to 'groupchat_data'.
    • custom_df (pandas.DataFrame): A DataFrame for creating a custom table in storage. Defaults to None.

Private Methods:

_ensure_storage_directory_exists(self)
  • Purpose: Ensures the existence of the storage directory, creating it if necessary.
_load_state(self)
  • Purpose: Loads the chat context from a file.
  • Returns: A dictionary representing the loaded chat context.

Public Methods:

persist(self)
  • Purpose: Persists the current chat context to a file.
add_message(self, sender_id, message, additional_info=None)
  • Purpose: Adds a message to the chat storage.
  • Parameters:
    • sender_id (str): ID of the message sender.
    • message (str): Content of the message.
    • additional_info (dict, optional): Additional message information. Defaults to None.
add_messages(self, messages)
  • Purpose: Adds multiple messages to the chat storage.
  • Parameters:
    • messages (pandas.DataFrame or list): Messages to be added.
remove_message(self, message_id)
  • Purpose: Removes a message from the chat storage.
  • Parameters:
    • message_id (int): ID of the message to be removed.
get_messages(self, type='dataframe')
  • Purpose: Retrieves messages from the chat storage.
  • Parameters:
    • type (str): Type of the returned messages ('dataframe' or 'json'). Defaults to 'dataframe'.
  • Returns: pandas.DataFrame or list of messages.
get_current_context(self)
  • Purpose: Retrieves the current chat context.
  • Returns: A dictionary representing the current chat context.
update_context(self, new_context)
  • Purpose: Updates the chat context with new values.
  • Parameters:
    • new_context (dict): New chat context values.
add_agent(self, agent)
  • Purpose: Adds an agent to the chat.
  • Parameters:
    • agent (Agent): The agent to be added.
remove_agent(self, agent_id)
  • Purpose: Removes an agent from the chat.
  • Parameters:
    • agent_id (str): ID of the agent to be removed.

Private Helper Methods:

_messages_to_dataframe(self, messages)
  • Purpose: Converts messages to a pandas DataFrame.
  • Parameters:
    • messages (list): Messages to be converted.
  • Returns: pandas.DataFrame of messages.
_messages_to_json(self, messages)
  • Purpose: Converts messages to JSON format.
  • Parameters:
    • messages (list): Messages to be converted.
  • Returns: list of messages in JSON format.

Example Usage

Initializing a Chat Session

chat_session = Chat(storage_path='my_chat_data')

Adding a Message

chat_session.add_message(sender_id='user123', message='Hello, world!')

Retrieving Messages as DataFrame

messages_df = chat_session.get_messages(type='dataframe')

Updating Chat Context

new_context = {'topic': 'AI'}
chat_session.update_context(new_context)

Adding an Agent

Assuming agent is an instance of the Agent class:

chat_session.add_agent(agent)

Notes

  • Ensure that the storage path provided during initialization is accessible and writable.
  • The Chat class is designed to interact seamlessly with other components of the MiniAutoGen framework, such as Agent and ChatStorage.