autogen.logger.base_logger.BaseLogger

BaseLogger

BaseLogger()

Helper class that provides a standard way to create an ABC using inheritance.

Instance Methods

get_connection

get_connection(self) -> None | Connection

Return a connection to the logging database.

log_chat_completion

log_chat_completion(
    self,
    invocation_id: UUID,
    client_id: int,
    wrapper_id: int,
    source: str | Agent,
    request: dict[str, float | str | list[dict[str, str]]],
    response: str | ChatCompletion,
    is_cached: int,
    cost: float,
    start_time: str
) -> None

Log a chat completion to database.
In AG2, chat completions are somewhat complicated because they are handled by the autogen.oai.OpenAIWrapper class.
One invocation to create can lead to multiple underlying OpenAI calls, depending on the llm_config list used, and any errors or retries.

Parameters:

Name	Description
`invocation_id`	A unique identifier for the invocation to the OpenAIWrapper.create method call Type: uuid.UUID
`client_id`	A unique identifier for the underlying OpenAI client instance Type: int
`wrapper_id`	A unique identifier for the OpenAIWrapper instance Type: int
`source`	The source/creator of the event as a string name or an Agent instance Type: str \| Agent
`request`	A dictionary representing the request or call to the OpenAI client endpoint Type: dict[str, float \| str \| list[dict[str, str]]]
`response`	The response from OpenAI Type: str \| ChatCompletion
`is_cached`	1 if the response was a cache hit, 0 otherwise Type: int
`cost`	The cost for OpenAI response Type: float
`start_time`	A string representing the moment the request was initiated Type: str

log_event

log_event(
    self,
    source: str | Agent,
    name: str,
    **kwargs: dict[str, Any]
) -> None

Log an event for an agent.

Parameters:

Name	Description
`source`	The source/creator of the event as a string name or an Agent instance Type: str \| Agent
`name`	The name of the event Type: str
`**kwargs`	Type: dict[str, Any]

log_function_use

log_function_use(
    self,
    source: str | Agent,
    function: F,
    args: dict[str, Any],
    returns: Any
) -> None

Log the use of a registered function (could be a tool)

Parameters:

Name	Description
`source`	The source/creator of the event as a string name or an Agent instance Type: str \| Agent
`function`	The function information Type: F
`args`	The function args to log Type: dict[str, Any]
`returns`	The return Type: Any

log_new_agent

log_new_agent(
    self,
    agent: ConversableAgent,
    init_args: dict[str, Any]
) -> None

Log the birth of a new agent.

Parameters:

Name	Description
`agent`	The agent to log. Type: ConversableAgent
`init_args`	The arguments passed to the construct the conversable agent Type: dict[str, Any]

log_new_client

log_new_client(
    self,
    client: AzureOpenAI | OpenAI,
    wrapper: OpenAIWrapper,
    init_args: dict[str, Any]
) -> None

Log the birth of a new OpenAIWrapper.

Parameters:

Name	Description
`client`	Type: AzureOpenAI \| OpenAI
`wrapper`	The OpenAI client to log. Type: OpenAIWrapper
`init_args`	The arguments passed to the construct the client Type: dict[str, Any]

log_new_wrapper

log_new_wrapper(
    self,
    wrapper: OpenAIWrapper,
    init_args: dict[str, LLMConfig | list[LLMConfig]]
) -> None

Log the birth of a new OpenAIWrapper.

Parameters:

Name	Description
`wrapper`	The wrapper to log. Type: OpenAIWrapper
`init_args`	The arguments passed to the construct the wrapper Type: dict[str, LLMConfig \| list[LLMConfig]]

start

start(self) -> str

Open a connection to the logging database, and start recording.
Returns:
session_id (str): a unique id for the logging session

Returns:

Type	Description
str	session_id (str): a unique id for the logging session

stop

stop(self) -> None

Close the connection to the logging database, and stop logging.

API Reference

​BaseLogger

​Instance Methods

​get_connection

​log_chat_completion

​log_event

​log_function_use

​log_new_agent

​log_new_client

​log_new_wrapper

​start

​stop

BaseLogger

Instance Methods

get_connection

log_chat_completion

log_event

log_function_use

log_new_agent

log_new_client

log_new_wrapper

start

stop