camel.societies package

Submodules

camel.societies.babyagi_playing module

class camel.societies.babyagi_playing.BabyAGI(assistant_role_name: str, user_role_name: str, task_prompt: str = '', task_type: TaskType = TaskType.AI_SOCIETY, max_task_history: int = 10, assistant_agent_kwargs: Dict | None = None, task_specify_agent_kwargs: Dict | None = None, task_creation_agent_kwargs: Dict | None = None, task_prioritization_agent_kwargs: Dict | None = None, sys_msg_generator_kwargs: Dict | None = None, extend_task_specify_meta_dict: Dict | None = None, output_language: str | None = None, message_window_size: int | None = None)

Bases: object

The BabyAGI Agent adapted from “Task-driven Autonomous Agent”.

Parameters:

• assistant_role_name (str) – The name of the role played by the assistant.

• user_role_name (str) – The name of the role played by the user.

• task_prompt (str, optional) – A prompt for the task to be performed. (default: "")

• task_type (TaskType, optional) – The type of task to perform. (default: TaskType.AI_SOCIETY)

• max_task_history (int) – The maximum number of previous tasks information to include in the task agent. (default: :obj:10)

• assistant_agent_kwargs (Dict, optional) – Additional arguments to pass to the assistant agent. (default: None)

• task_specify_agent_kwargs (Dict, optional) – Additional arguments to pass to the task specify agent. (default: None)

• task_creation_agent_kwargs (Dict, optional) – Additional arguments to pass to the task creation agent. (default: None)

• task_prioritization_agent_kwargs (Dict, optional) – Additional arguments to pass to the task prioritization agent. (default: None)

• sys_msg_generator_kwargs (Dict, optional) – Additional arguments to pass to the system message generator. (default: None)

• extend_task_specify_meta_dict (Dict, optional) – A dict to extend the task specify meta dict with. (default: None)

• output_language (str, optional) – The language to be output by the agents. (default: None)

• message_window_size (int, optional) – The maximum number of previous messages to include in the context window. If None, no windowing is performed. (default: None)

init_agents(init_assistant_sys_msg: BaseMessage, assistant_agent_kwargs: Dict | None, task_creation_agent_kwargs: Dict | None, task_prioritization_agent_kwargs: Dict | None, output_language: str | None, message_window_size: int | None = None)

Initialize assistant and user agents with their system messages.

Parameters:

• init_assistant_sys_msg (BaseMessage) – Assistant agent’s initial system message.

• assistant_agent_kwargs (Dict, optional) – Additional arguments to pass to the assistant agent.

• task_creation_agent_kwargs (Dict, optional) – Additional arguments to pass to the task creation agent.

• task_prioritization_agent_kwargs (Dict, optional) – Additional arguments to pass to the task prioritization agent.

• output_language (str, optional) – The language to be output by the agents.

• message_window_size (int, optional) – The maximum number of previous messages to include in the context window. If None, no windowing is performed. (default: None)

init_specified_task_prompt(assistant_role_name: str, user_role_name: str, task_specify_agent_kwargs: Dict | None, extend_task_specify_meta_dict: Dict | None, output_language: str | None)

Use a task specify agent to generate a specified task prompt. Generated specified task prompt will be used to replace original task prompt. If there is no task specify agent, specified task prompt will not be generated.

Parameters:

• assistant_role_name (str) – The name of the role played by the assistant.

• user_role_name (str) – The name of the role played by the user.

• task_specify_agent_kwargs (Dict, optional) – Additional arguments to pass to the task specify agent.

• extend_task_specify_meta_dict (Dict, optional) – A dict to extend the task specify meta dict with.

• output_language (str, optional) – The language to be output by the agents.

step() → ChatAgentResponse

BabyAGI agent would pull the first task from the task list, complete the task based on the context, then creates new tasks and re-prioritizes the task list based on the objective and the result of the previous task. It returns assistant message.

Returns:

it contains the resulting assistant message, whether the assistant agent terminated the conversation, and any additional assistant information.

Return type:

ChatAgentResponse

camel.societies.role_playing module

class camel.societies.role_playing.RolePlaying(assistant_role_name: str, user_role_name: str, *, critic_role_name: str = 'critic', task_prompt: str = '', with_task_specify: bool = True, with_task_planner: bool = False, with_critic_in_the_loop: bool = False, critic_criteria: str | None = None, model: BaseModelBackend | None = None, task_type: TaskType = TaskType.AI_SOCIETY, assistant_agent_kwargs: Dict | None = None, user_agent_kwargs: Dict | None = None, task_specify_agent_kwargs: Dict | None = None, task_planner_agent_kwargs: Dict | None = None, critic_kwargs: Dict | None = None, sys_msg_generator_kwargs: Dict | None = None, extend_sys_msg_meta_dicts: List[Dict] | None = None, extend_task_specify_meta_dict: Dict | None = None, output_language: str | None = None)

Bases: object

Role playing between two agents.

Parameters:

• assistant_role_name (str) – The name of the role played by the assistant.

• user_role_name (str) – The name of the role played by the user.

• critic_role_name (str, optional) – The name of the role played by the critic. Role name with "human" will set critic as a Human agent, else will create a CriticAgent. (default: "critic")

• task_prompt (str, optional) – A prompt for the task to be performed. (default: "")

• with_task_specify (bool, optional) – Whether to use a task specify agent. (default: True)

• with_task_planner (bool, optional) – Whether to use a task planner agent. (default: False)

• with_critic_in_the_loop (bool, optional) – Whether to include a critic in the loop. (default: False)

• critic_criteria (str, optional) – Critic criteria for the critic agent. If not specified, set the criteria to improve task performance.

• model (BaseModelBackend, optional) – The model backend to use for generating responses. If specified, it will override the model in all agents if not specified in agent-specific kwargs. (default: OpenAIModel with GPT_4O_MINI)

• task_type (TaskType, optional) – The type of task to perform. (default: TaskType.AI_SOCIETY)

• assistant_agent_kwargs (Dict, optional) – Additional arguments to pass to the assistant agent. (default: None)

• user_agent_kwargs (Dict, optional) – Additional arguments to pass to the user agent. (default: None)

• task_specify_agent_kwargs (Dict, optional) – Additional arguments to pass to the task specify agent. (default: None)

• task_planner_agent_kwargs (Dict, optional) – Additional arguments to pass to the task planner agent. (default: None)

• critic_kwargs (Dict, optional) – Additional arguments to pass to the critic. (default: None)

• sys_msg_generator_kwargs (Dict, optional) – Additional arguments to pass to the system message generator. (default: None)

• extend_sys_msg_meta_dicts (List[Dict], optional) – A list of dicts to extend the system message meta dicts with. (default: None)

• extend_task_specify_meta_dict (Dict, optional) – A dict to extend the task specify meta dict with. (default: None)

• output_language (str, optional) – The language to be output by the agents. (default: None)

init_chat(init_msg_content: str | None = None) → BaseMessage

Initializes the chat by resetting both of the assistant and user agents. Returns an initial message for the role-playing session.

Parameters:

init_msg_content (str, optional) – A user-specified initial message. Will be sent to the role-playing session as the initial message. (default: None)

Returns:

A single BaseMessage representing the initial

message.

Return type:

BaseMessage

step(assistant_msg: BaseMessage) → Tuple[ChatAgentResponse, ChatAgentResponse]

Advances the conversation by taking a message from the assistant, processing it using the user agent, and then processing the resulting message using the assistant agent. Returns a tuple containing the resulting assistant message, whether the assistant agent terminated the conversation, and any additional assistant information, as well as a tuple containing the resulting user message, whether the user agent terminated the conversation, and any additional user information.

Parameters:

assistant_msg – A BaseMessage representing the message from the assistant.

Returns:

A tuple containing two

ChatAgentResponse: the first struct contains the resulting assistant message, whether the assistant agent terminated the conversation, and any additional assistant information; the second struct contains the resulting user message, whether the user agent terminated the conversation, and any additional user information.

Return type:

Tuple[ChatAgentResponse, ChatAgentResponse]

Subpackages

• camel.societies.workforce package
  • Submodules
  • camel.societies.workforce.base module
    • BaseNode
      • BaseNode.reset()
      • BaseNode.set_channel()
      • BaseNode.start()
      • BaseNode.stop()
  • camel.societies.workforce.prompts module
  • camel.societies.workforce.role_playing_worker module
    • RolePlayingWorker
  • camel.societies.workforce.single_agent_worker module
    • SingleAgentWorker
      • SingleAgentWorker.reset()
  • camel.societies.workforce.task_channel module
    • Packet
      • Packet.task
      • Packet.publisher_id
      • Packet.assignee_id
      • Packet.status
    • PacketStatus
      • PacketStatus.ARCHIVED
      • PacketStatus.RETURNED
      • PacketStatus.SENT
    • TaskChannel
      • TaskChannel.archive_task()
      • TaskChannel.get_assigned_task_by_assignee()
      • TaskChannel.get_channel_debug_info()
      • TaskChannel.get_dependency_ids()
      • TaskChannel.get_returned_task_by_publisher()
      • TaskChannel.get_task_by_id()
      • TaskChannel.post_dependency()
      • TaskChannel.post_task()
      • TaskChannel.remove_task()
      • TaskChannel.return_task()
  • camel.societies.workforce.utils module
    • TaskAssignResult
      • TaskAssignResult.assignee_id
      • TaskAssignResult.model_computed_fields
      • TaskAssignResult.model_config
      • TaskAssignResult.model_fields
    • TaskResult
      • TaskResult.content
      • TaskResult.failed
      • TaskResult.model_computed_fields
      • TaskResult.model_config
      • TaskResult.model_fields
    • WorkerConf
      • WorkerConf.description
      • WorkerConf.model_computed_fields
      • WorkerConf.model_config
      • WorkerConf.model_fields
      • WorkerConf.role
      • WorkerConf.sys_msg
    • check_if_running()
  • camel.societies.workforce.worker module
    • Worker
      • Worker.set_channel()
      • Worker.start()
      • Worker.stop()
  • camel.societies.workforce.workforce module
    • Workforce
      • Workforce.add_role_playing_worker()
      • Workforce.add_single_agent_worker()
      • Workforce.add_workforce()
      • Workforce.process_task()
      • Workforce.reset()
      • Workforce.set_channel()
      • Workforce.start()
      • Workforce.stop()
  • Module contents
    • RolePlayingWorker
    • SingleAgentWorker
      • SingleAgentWorker.reset()
    • Workforce
      • Workforce.add_role_playing_worker()
      • Workforce.add_single_agent_worker()
      • Workforce.add_workforce()
      • Workforce.process_task()
      • Workforce.reset()
      • Workforce.set_channel()
      • Workforce.start()
      • Workforce.stop()

Module contents

class camel.societies.BabyAGI(assistant_role_name: str, user_role_name: str, task_prompt: str = '', task_type: TaskType = TaskType.AI_SOCIETY, max_task_history: int = 10, assistant_agent_kwargs: Dict | None = None, task_specify_agent_kwargs: Dict | None = None, task_creation_agent_kwargs: Dict | None = None, task_prioritization_agent_kwargs: Dict | None = None, sys_msg_generator_kwargs: Dict | None = None, extend_task_specify_meta_dict: Dict | None = None, output_language: str | None = None, message_window_size: int | None = None)

Bases: object

The BabyAGI Agent adapted from “Task-driven Autonomous Agent”.

Parameters:

• assistant_role_name (str) – The name of the role played by the assistant.

• user_role_name (str) – The name of the role played by the user.

• task_prompt (str, optional) – A prompt for the task to be performed. (default: "")

• task_type (TaskType, optional) – The type of task to perform. (default: TaskType.AI_SOCIETY)

• max_task_history (int) – The maximum number of previous tasks information to include in the task agent. (default: :obj:10)

• assistant_agent_kwargs (Dict, optional) – Additional arguments to pass to the assistant agent. (default: None)

• task_specify_agent_kwargs (Dict, optional) – Additional arguments to pass to the task specify agent. (default: None)

• task_creation_agent_kwargs (Dict, optional) – Additional arguments to pass to the task creation agent. (default: None)

• task_prioritization_agent_kwargs (Dict, optional) – Additional arguments to pass to the task prioritization agent. (default: None)

• sys_msg_generator_kwargs (Dict, optional) – Additional arguments to pass to the system message generator. (default: None)

• extend_task_specify_meta_dict (Dict, optional) – A dict to extend the task specify meta dict with. (default: None)

• output_language (str, optional) – The language to be output by the agents. (default: None)

• message_window_size (int, optional) – The maximum number of previous messages to include in the context window. If None, no windowing is performed. (default: None)

init_agents(init_assistant_sys_msg: BaseMessage, assistant_agent_kwargs: Dict | None, task_creation_agent_kwargs: Dict | None, task_prioritization_agent_kwargs: Dict | None, output_language: str | None, message_window_size: int | None = None)

Initialize assistant and user agents with their system messages.

Parameters:

• init_assistant_sys_msg (BaseMessage) – Assistant agent’s initial system message.

• assistant_agent_kwargs (Dict, optional) – Additional arguments to pass to the assistant agent.

• task_creation_agent_kwargs (Dict, optional) – Additional arguments to pass to the task creation agent.

• task_prioritization_agent_kwargs (Dict, optional) – Additional arguments to pass to the task prioritization agent.

• output_language (str, optional) – The language to be output by the agents.

• message_window_size (int, optional) – The maximum number of previous messages to include in the context window. If None, no windowing is performed. (default: None)

init_specified_task_prompt(assistant_role_name: str, user_role_name: str, task_specify_agent_kwargs: Dict | None, extend_task_specify_meta_dict: Dict | None, output_language: str | None)

Use a task specify agent to generate a specified task prompt. Generated specified task prompt will be used to replace original task prompt. If there is no task specify agent, specified task prompt will not be generated.

Parameters:

• assistant_role_name (str) – The name of the role played by the assistant.

• user_role_name (str) – The name of the role played by the user.

• task_specify_agent_kwargs (Dict, optional) – Additional arguments to pass to the task specify agent.

• extend_task_specify_meta_dict (Dict, optional) – A dict to extend the task specify meta dict with.

• output_language (str, optional) – The language to be output by the agents.

step() → ChatAgentResponse

BabyAGI agent would pull the first task from the task list, complete the task based on the context, then creates new tasks and re-prioritizes the task list based on the objective and the result of the previous task. It returns assistant message.

Returns:

it contains the resulting assistant message, whether the assistant agent terminated the conversation, and any additional assistant information.

Return type:

ChatAgentResponse

class camel.societies.RolePlaying(assistant_role_name: str, user_role_name: str, *, critic_role_name: str = 'critic', task_prompt: str = '', with_task_specify: bool = True, with_task_planner: bool = False, with_critic_in_the_loop: bool = False, critic_criteria: str | None = None, model: BaseModelBackend | None = None, task_type: TaskType = TaskType.AI_SOCIETY, assistant_agent_kwargs: Dict | None = None, user_agent_kwargs: Dict | None = None, task_specify_agent_kwargs: Dict | None = None, task_planner_agent_kwargs: Dict | None = None, critic_kwargs: Dict | None = None, sys_msg_generator_kwargs: Dict | None = None, extend_sys_msg_meta_dicts: List[Dict] | None = None, extend_task_specify_meta_dict: Dict | None = None, output_language: str | None = None)

Bases: object

Role playing between two agents.

Parameters:

• assistant_role_name (str) – The name of the role played by the assistant.

• user_role_name (str) – The name of the role played by the user.

• critic_role_name (str, optional) – The name of the role played by the critic. Role name with "human" will set critic as a Human agent, else will create a CriticAgent. (default: "critic")

• task_prompt (str, optional) – A prompt for the task to be performed. (default: "")

• with_task_specify (bool, optional) – Whether to use a task specify agent. (default: True)

• with_task_planner (bool, optional) – Whether to use a task planner agent. (default: False)

• with_critic_in_the_loop (bool, optional) – Whether to include a critic in the loop. (default: False)

• critic_criteria (str, optional) – Critic criteria for the critic agent. If not specified, set the criteria to improve task performance.

• model (BaseModelBackend, optional) – The model backend to use for generating responses. If specified, it will override the model in all agents if not specified in agent-specific kwargs. (default: OpenAIModel with GPT_4O_MINI)

• task_type (TaskType, optional) – The type of task to perform. (default: TaskType.AI_SOCIETY)

• assistant_agent_kwargs (Dict, optional) – Additional arguments to pass to the assistant agent. (default: None)

• user_agent_kwargs (Dict, optional) – Additional arguments to pass to the user agent. (default: None)

• task_specify_agent_kwargs (Dict, optional) – Additional arguments to pass to the task specify agent. (default: None)

• task_planner_agent_kwargs (Dict, optional) – Additional arguments to pass to the task planner agent. (default: None)

• critic_kwargs (Dict, optional) – Additional arguments to pass to the critic. (default: None)

• sys_msg_generator_kwargs (Dict, optional) – Additional arguments to pass to the system message generator. (default: None)

• extend_sys_msg_meta_dicts (List[Dict], optional) – A list of dicts to extend the system message meta dicts with. (default: None)

• extend_task_specify_meta_dict (Dict, optional) – A dict to extend the task specify meta dict with. (default: None)

• output_language (str, optional) – The language to be output by the agents. (default: None)

init_chat(init_msg_content: str | None = None) → BaseMessage

Initializes the chat by resetting both of the assistant and user agents. Returns an initial message for the role-playing session.

Parameters:

init_msg_content (str, optional) – A user-specified initial message. Will be sent to the role-playing session as the initial message. (default: None)

Returns:

A single BaseMessage representing the initial

message.

Return type:

BaseMessage

step(assistant_msg: BaseMessage) → Tuple[ChatAgentResponse, ChatAgentResponse]

Advances the conversation by taking a message from the assistant, processing it using the user agent, and then processing the resulting message using the assistant agent. Returns a tuple containing the resulting assistant message, whether the assistant agent terminated the conversation, and any additional assistant information, as well as a tuple containing the resulting user message, whether the user agent terminated the conversation, and any additional user information.

Parameters:

assistant_msg – A BaseMessage representing the message from the assistant.

Returns:

A tuple containing two

ChatAgentResponse: the first struct contains the resulting assistant message, whether the assistant agent terminated the conversation, and any additional assistant information; the second struct contains the resulting user message, whether the user agent terminated the conversation, and any additional user information.

Return type:

Tuple[ChatAgentResponse, ChatAgentResponse]