latent.mlflow.wrapper¶

Safe MLFlow wrapper that automatically handles active run checks.

This module provides the same interface as mlflow but gracefully handles cases where no active run exists (e.g., when tasks execute outside the flow's MLFlow context).

Usage: from evals.prefect_utils import mlflow

mlflow.log_metric("key", value)  # Safe - checks for active run
mlflow.log_param("key", value)   # Safe - checks for active run

Attributes¶

MlflowClient¶

active_run¶

end_run¶

get_experiment¶

get_experiment_by_name¶

langchain¶

litellm¶

log_artifact¶

log_artifacts¶

log_metric¶

log_metrics¶

log_param¶

log_params¶

set_experiment¶

set_tag¶

set_tags¶

set_tracking_uri¶

start_run¶

start_span¶