CAMEL - Build Multi-Agent AI Systems

JupyterKernelInterpreter

class JupyterKernelInterpreter(BaseInterpreter):

A class for executing code strings in a Jupyter Kernel. Parameters:

require_confirm (bool, optional): If True, prompt user before running code strings for security. Defaults to True.
print_stdout (bool, optional): If True, print the standard output of the executed code. Defaults to False.
print_stderr (bool, optional): If True, print the standard error of the executed code. Defaults to True.

def __init__(
    self,
    require_confirm: bool = True,
    print_stdout: bool = False,
    print_stderr: bool = True
):

def __del__(self):

Clean up the kernel and client.

def _initialize_if_needed(self):

Initialize the kernel manager and client if they are not already initialized.

def _clean_ipython_output(output: str):

Remove ANSI escape sequences from the output.

def _execute(self, code: str, timeout: float):

Execute the code in the Jupyter kernel and return the result.

def run(self, code: str, code_type: str = 'python'):

Executes the given code in the Jupyter kernel. Parameters:

code (str): The code string to execute.
code_type (str): The type of code to execute (e.g., ‘python’, ‘bash’). (default: obj:python)

Returns: str: A string containing the captured result of the executed code.

def execute_command(self, command: str):

Executes a shell command in the Jupyter kernel. Parameters:

Returns: str: A string containing the captured result of the executed command.

def supported_code_types(self):

Returns: List[str]: Supported code types.

def update_action_space(self, action_space: Dict[str, Any]):

Updates the action space for the interpreter. Parameters:

action_space (Dict[str, Any]): A dictionary representing the new or updated action space.