MlflowCallbackHandler#

class langchain_community.callbacks.mlflow_callback.MlflowCallbackHandler(name: str | None = 'langchainrun-%', experiment: str | None = 'langchain', tags: Dict | None = None, tracking_uri: str | None = None, run_id: str | None = None, artifacts_dir: str = '')[source]#

Callback Handler that logs metrics and artifacts to mlflow server.

Parameters:

name (str) – Name of the run.
experiment (str) – Name of the experiment.
tags (dict) – Tags to be attached for the run.
tracking_uri (str) – MLflow tracking server uri.
run_id (str | None)
artifacts_dir (str)

This handler will utilize the associated callback method called and formats the input of each callback function with metadata regarding the state of LLM run, and adds the response to the list of records for both the {method}_records and action. It then logs the response to mlflow server.

Initialize callback handler.

Attributes

`always_verbose`	Whether to call verbose callbacks even if verbose is False.
`ignore_agent`	Whether to ignore agent callbacks.
`ignore_chain`	Whether to ignore chain callbacks.
`ignore_chat_model`	Whether to ignore chat model callbacks.
`ignore_custom_event`	Ignore custom event.
`ignore_llm`	Whether to ignore LLM callbacks.
`ignore_retriever`	Whether to ignore retriever callbacks.
`ignore_retry`	Whether to ignore retry callbacks.
`raise_error`	Whether to raise an error if an exception occurs.
`run_inline`	Whether to run the callback inline.

Methods

`__init__`([name, experiment, tags, ...])	Initialize callback handler.
`flush_tracker`([langchain_asset, finish])
`get_custom_callback_meta`()
`on_agent_action`(action, **kwargs)	Run on agent action.
`on_agent_finish`(finish, **kwargs)	Run when agent ends running.
`on_chain_end`(outputs, **kwargs)	Run when chain ends running.
`on_chain_error`(error, **kwargs)	Run when chain errors.
`on_chain_start`(serialized, inputs, **kwargs)	Run when chain starts running.
`on_chat_model_start`(serialized, messages, *, ...)	Run when a chat model starts running.
`on_custom_event`(name, data, *, run_id[, ...])	Override to define a handler for a custom event.
`on_llm_end`(response, **kwargs)	Run when LLM ends running.
`on_llm_error`(error, **kwargs)	Run when LLM errors.
`on_llm_new_token`(token, **kwargs)	Run when LLM generates a new token.
`on_llm_start`(serialized, prompts, **kwargs)	Run when LLM starts.
`on_retriever_end`(documents, **kwargs)	Run when Retriever ends running.
`on_retriever_error`(error, **kwargs)	Run when Retriever errors.
`on_retriever_start`(serialized, query, **kwargs)	Run when Retriever starts running.
`on_retry`(retry_state, *, run_id[, parent_run_id])	Run on a retry event.
`on_text`(text, **kwargs)	Run when text is received.
`on_tool_end`(output, **kwargs)	Run when tool ends running.
`on_tool_error`(error, **kwargs)	Run when tool errors.
`on_tool_start`(serialized, input_str, **kwargs)	Run when tool starts running.
`reset_callback_meta`()	Reset the callback metadata.

Initialize callback handler.

Parameters:

name (str | None)
experiment (str | None)
tags (Dict | None)
tracking_uri (str | None)
run_id (str | None)
artifacts_dir (str)

Return type:

None

flush_tracker(langchain_asset: Any = None, finish: bool = False) → None[source]#

Parameters:

langchain_asset (Any)
finish (bool)

Return type:

None

get_custom_callback_meta() → Dict[str, Any]#

Return type:: Dict[str, Any]

on_agent_action(action: AgentAction, **kwargs: Any) → Any[source]#

Run on agent action.

Parameters:

action (AgentAction)
kwargs (Any)

Return type:

Any

on_agent_finish(finish: AgentFinish, **kwargs: Any) → None[source]#

Run when agent ends running.

Parameters:

finish (AgentFinish)
kwargs (Any)

Return type:

None

on_chain_end(outputs: Dict[str, Any] | str | List[str], **kwargs: Any) → None[source]#

Run when chain ends running.

Parameters:

outputs (Dict[str, Any] | str | List[str])
kwargs (Any)

Return type:

None

on_chain_error(error: BaseException, **kwargs: Any) → None[source]#

Run when chain errors.

Parameters:

error (BaseException)
kwargs (Any)

Return type:

None

on_chain_start(serialized: Dict[str, Any], inputs: Dict[str, Any], **kwargs: Any) → None[source]#

Run when chain starts running.

Parameters:

serialized (Dict[str, Any])
inputs (Dict[str, Any])
kwargs (Any)

Return type:

None

on_chat_model_start(serialized: dict[str, Any], messages: list[list[BaseMessage]], *, run_id: UUID, parent_run_id: UUID | None = None, tags: list[str] | None = None, metadata: dict[str, Any] | None = None, **kwargs: Any) → Any#

Run when a chat model starts running.

ATTENTION: This method is called for chat models. If you’re implementing: a handler for a non-chat model, you should use on_llm_start instead.

Parameters:

serialized (Dict[str, Any]) – The serialized chat model.
messages (List[List[BaseMessage]]) – The messages.
run_id (UUID) – The run ID. This is the ID of the current run.
parent_run_id (UUID) – The parent run ID. This is the ID of the parent run.
tags (Optional[List[str]]) – The tags.
metadata (Optional[Dict[str, Any]]) – The metadata.
kwargs (Any) – Additional keyword arguments.

Return type:

Any

on_custom_event(name: str, data: Any, *, run_id: UUID, tags: list[str] | None = None, metadata: dict[str, Any] | None = None, **kwargs: Any) → Any#

Override to define a handler for a custom event.

Parameters:

name (str) – The name of the custom event.
data (Any) – The data for the custom event. Format will match the format specified by the user.
run_id (UUID) – The ID of the run.
tags (list[str] | None) – The tags associated with the custom event (includes inherited tags).
metadata (dict[str, Any] | None) – The metadata associated with the custom event (includes inherited metadata).
kwargs (Any)

Return type:

Any

Added in version 0.2.15.

on_llm_end(response: LLMResult, **kwargs: Any) → None[source]#

Run when LLM ends running.

Parameters:

response (LLMResult)
kwargs (Any)

Return type:

None

on_llm_error(error: BaseException, **kwargs: Any) → None[source]#

Run when LLM errors.

Parameters:

error (BaseException)
kwargs (Any)

Return type:

None

on_llm_new_token(token: str, **kwargs: Any) → None[source]#

Run when LLM generates a new token.

Parameters:

token (str)
kwargs (Any)

Return type:

None

on_llm_start(serialized: Dict[str, Any], prompts: List[str], **kwargs: Any) → None[source]#

Run when LLM starts.

Parameters:

serialized (Dict[str, Any])
prompts (List[str])
kwargs (Any)

Return type:

None

on_retriever_end(documents: Sequence[Document], **kwargs: Any) → Any[source]#

Run when Retriever ends running.

Parameters:

documents (Sequence[Document])
kwargs (Any)

Return type:

Any

on_retriever_error(error: BaseException, **kwargs: Any) → Any[source]#

Run when Retriever errors.

Parameters:

error (BaseException)
kwargs (Any)

Return type:

Any

on_retriever_start(serialized: Dict[str, Any], query: str, **kwargs: Any) → Any[source]#

Run when Retriever starts running.

Parameters:

serialized (Dict[str, Any])
query (str)
kwargs (Any)

Return type:

Any

on_retry(retry_state: RetryCallState, *, run_id: UUID, parent_run_id: UUID | None = None, **kwargs: Any) → Any#

Run on a retry event.

Parameters:

retry_state (RetryCallState) – The retry state.
run_id (UUID) – The run ID. This is the ID of the current run.
parent_run_id (UUID) – The parent run ID. This is the ID of the parent run.
kwargs (Any) – Additional keyword arguments.

Return type:

Any

on_text(text: str, **kwargs: Any) → None[source]#

Run when text is received.

Parameters:

text (str)
kwargs (Any)

Return type:

None

on_tool_end(output: Any, **kwargs: Any) → None[source]#

Run when tool ends running.

Parameters:

output (Any)
kwargs (Any)

Return type:

None

on_tool_error(error: BaseException, **kwargs: Any) → None[source]#

Run when tool errors.

Parameters:

error (BaseException)
kwargs (Any)

Return type:

None

on_tool_start(serialized: Dict[str, Any], input_str: str, **kwargs: Any) → None[source]#

Run when tool starts running.

Parameters:

serialized (Dict[str, Any])
input_str (str)
kwargs (Any)

Return type:

None

reset_callback_meta() → None#

Reset the callback metadata.

Return type:: None

Examples using MlflowCallbackHandler

MLflow